RAG_Test1.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "provenance": [],
      "machine_shape": "hm",
      "gpuType": "T4",
      "authorship_tag": "ABX9TyPrjHpAtayIF3RcLxZEAwf/",
      "include_colab_link": true
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    },
    "language_info": {
      "name": "python"
    },
    "accelerator": "GPU",
    "widgets": {
      "application/vnd.jupyter.widget-state+json": {
        "01a5294ce7d04a2f93d9169e26f4a508": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "HBoxModel",
          "model_module_version": "1.5.0",
          "state": {
            "_dom_classes": [],
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "HBoxModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/controls",
            "_view_module_version": "1.5.0",
            "_view_name": "HBoxView",
            "box_style": "",
            "children": [
              "IPY_MODEL_f1bf2f7393804ab98858767a04530e4d",
              "IPY_MODEL_34628447104e4a44a139ddb705b02150",
              "IPY_MODEL_cfd8aa90600241bd97f3161bce9763a6"
            ],
            "layout": "IPY_MODEL_5c9db98af10f4b3f8f746dab51b6234d"
          }
        },
        "f1bf2f7393804ab98858767a04530e4d": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "HTMLModel",
          "model_module_version": "1.5.0",
          "state": {
            "_dom_classes": [],
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "HTMLModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/controls",
            "_view_module_version": "1.5.0",
            "_view_name": "HTMLView",
            "description": "",
            "description_tooltip": null,
            "layout": "IPY_MODEL_af0a9644c4174d809af0ef4d3dfa7a0c",
            "placeholder": "​",
            "style": "IPY_MODEL_001856a52d2b4e0e9fd9cb674a329dd5",
            "value": ""
          }
        },
        "34628447104e4a44a139ddb705b02150": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "FloatProgressModel",
          "model_module_version": "1.5.0",
          "state": {
            "_dom_classes": [],
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "FloatProgressModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/controls",
            "_view_module_version": "1.5.0",
            "_view_name": "ProgressView",
            "bar_style": "success",
            "description": "",
            "description_tooltip": null,
            "layout": "IPY_MODEL_2c9f5b3371134d0eb9d357f14f576c93",
            "max": 1,
            "min": 0,
            "orientation": "horizontal",
            "style": "IPY_MODEL_9244e5e7a2ce4a3ebceaf4b21bf60c58",
            "value": 1
          }
        },
        "cfd8aa90600241bd97f3161bce9763a6": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "HTMLModel",
          "model_module_version": "1.5.0",
          "state": {
            "_dom_classes": [],
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "HTMLModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/controls",
            "_view_module_version": "1.5.0",
            "_view_name": "HTMLView",
            "description": "",
            "description_tooltip": null,
            "layout": "IPY_MODEL_b4452006845d4231b4d56a753b44b0f1",
            "placeholder": "​",
            "style": "IPY_MODEL_f885417eb70e4e7b9bb92452de2d8a5a",
            "value": " 11/? [00:00&lt;00:00, 181.08it/s]"
          }
        },
        "5c9db98af10f4b3f8f746dab51b6234d": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "model_module_version": "1.2.0",
          "state": {
            "_model_module": "@jupyter-widgets/base",
            "_model_module_version": "1.2.0",
            "_model_name": "LayoutModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "LayoutView",
            "align_content": null,
            "align_items": null,
            "align_self": null,
            "border": null,
            "bottom": null,
            "display": null,
            "flex": null,
            "flex_flow": null,
            "grid_area": null,
            "grid_auto_columns": null,
            "grid_auto_flow": null,
            "grid_auto_rows": null,
            "grid_column": null,
            "grid_gap": null,
            "grid_row": null,
            "grid_template_areas": null,
            "grid_template_columns": null,
            "grid_template_rows": null,
            "height": null,
            "justify_content": null,
            "justify_items": null,
            "left": null,
            "margin": null,
            "max_height": null,
            "max_width": null,
            "min_height": null,
            "min_width": null,
            "object_fit": null,
            "object_position": null,
            "order": null,
            "overflow": null,
            "overflow_x": null,
            "overflow_y": null,
            "padding": null,
            "right": null,
            "top": null,
            "visibility": null,
            "width": null
          }
        },
        "af0a9644c4174d809af0ef4d3dfa7a0c": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "model_module_version": "1.2.0",
          "state": {
            "_model_module": "@jupyter-widgets/base",
            "_model_module_version": "1.2.0",
            "_model_name": "LayoutModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "LayoutView",
            "align_content": null,
            "align_items": null,
            "align_self": null,
            "border": null,
            "bottom": null,
            "display": null,
            "flex": null,
            "flex_flow": null,
            "grid_area": null,
            "grid_auto_columns": null,
            "grid_auto_flow": null,
            "grid_auto_rows": null,
            "grid_column": null,
            "grid_gap": null,
            "grid_row": null,
            "grid_template_areas": null,
            "grid_template_columns": null,
            "grid_template_rows": null,
            "height": null,
            "justify_content": null,
            "justify_items": null,
            "left": null,
            "margin": null,
            "max_height": null,
            "max_width": null,
            "min_height": null,
            "min_width": null,
            "object_fit": null,
            "object_position": null,
            "order": null,
            "overflow": null,
            "overflow_x": null,
            "overflow_y": null,
            "padding": null,
            "right": null,
            "top": null,
            "visibility": null,
            "width": null
          }
        },
        "001856a52d2b4e0e9fd9cb674a329dd5": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "DescriptionStyleModel",
          "model_module_version": "1.5.0",
          "state": {
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "DescriptionStyleModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "StyleView",
            "description_width": ""
          }
        },
        "2c9f5b3371134d0eb9d357f14f576c93": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "model_module_version": "1.2.0",
          "state": {
            "_model_module": "@jupyter-widgets/base",
            "_model_module_version": "1.2.0",
            "_model_name": "LayoutModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "LayoutView",
            "align_content": null,
            "align_items": null,
            "align_self": null,
            "border": null,
            "bottom": null,
            "display": null,
            "flex": null,
            "flex_flow": null,
            "grid_area": null,
            "grid_auto_columns": null,
            "grid_auto_flow": null,
            "grid_auto_rows": null,
            "grid_column": null,
            "grid_gap": null,
            "grid_row": null,
            "grid_template_areas": null,
            "grid_template_columns": null,
            "grid_template_rows": null,
            "height": null,
            "justify_content": null,
            "justify_items": null,
            "left": null,
            "margin": null,
            "max_height": null,
            "max_width": null,
            "min_height": null,
            "min_width": null,
            "object_fit": null,
            "object_position": null,
            "order": null,
            "overflow": null,
            "overflow_x": null,
            "overflow_y": null,
            "padding": null,
            "right": null,
            "top": null,
            "visibility": null,
            "width": "20px"
          }
        },
        "9244e5e7a2ce4a3ebceaf4b21bf60c58": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "ProgressStyleModel",
          "model_module_version": "1.5.0",
          "state": {
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "ProgressStyleModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "StyleView",
            "bar_color": null,
            "description_width": ""
          }
        },
        "b4452006845d4231b4d56a753b44b0f1": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "model_module_version": "1.2.0",
          "state": {
            "_model_module": "@jupyter-widgets/base",
            "_model_module_version": "1.2.0",
            "_model_name": "LayoutModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "LayoutView",
            "align_content": null,
            "align_items": null,
            "align_self": null,
            "border": null,
            "bottom": null,
            "display": null,
            "flex": null,
            "flex_flow": null,
            "grid_area": null,
            "grid_auto_columns": null,
            "grid_auto_flow": null,
            "grid_auto_rows": null,
            "grid_column": null,
            "grid_gap": null,
            "grid_row": null,
            "grid_template_areas": null,
            "grid_template_columns": null,
            "grid_template_rows": null,
            "height": null,
            "justify_content": null,
            "justify_items": null,
            "left": null,
            "margin": null,
            "max_height": null,
            "max_width": null,
            "min_height": null,
            "min_width": null,
            "object_fit": null,
            "object_position": null,
            "order": null,
            "overflow": null,
            "overflow_x": null,
            "overflow_y": null,
            "padding": null,
            "right": null,
            "top": null,
            "visibility": null,
            "width": null
          }
        },
        "f885417eb70e4e7b9bb92452de2d8a5a": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "DescriptionStyleModel",
          "model_module_version": "1.5.0",
          "state": {
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "DescriptionStyleModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "StyleView",
            "description_width": ""
          }
        },
        "da8bf76154664474b489c8b1c118c732": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "HBoxModel",
          "model_module_version": "1.5.0",
          "state": {
            "_dom_classes": [],
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "HBoxModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/controls",
            "_view_module_version": "1.5.0",
            "_view_name": "HBoxView",
            "box_style": "",
            "children": [
              "IPY_MODEL_322abf6001f34aef8fd9aef4f3e4d9ef",
              "IPY_MODEL_d35d104d80f44614a3d5095e419a76ea",
              "IPY_MODEL_95d09829a7784ffb98b7b8012096c719"
            ],
            "layout": "IPY_MODEL_7b96a8f4b91c4dcf92c2d470b927cda5"
          }
        },
        "322abf6001f34aef8fd9aef4f3e4d9ef": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "HTMLModel",
          "model_module_version": "1.5.0",
          "state": {
            "_dom_classes": [],
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "HTMLModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/controls",
            "_view_module_version": "1.5.0",
            "_view_name": "HTMLView",
            "description": "",
            "description_tooltip": null,
            "layout": "IPY_MODEL_180ab15da4c545be9d48a7f358c3dfe2",
            "placeholder": "​",
            "style": "IPY_MODEL_92b34f04a0044697a9cdd2d617858cd8",
            "value": "100%"
          }
        },
        "d35d104d80f44614a3d5095e419a76ea": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "FloatProgressModel",
          "model_module_version": "1.5.0",
          "state": {
            "_dom_classes": [],
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "FloatProgressModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/controls",
            "_view_module_version": "1.5.0",
            "_view_name": "ProgressView",
            "bar_style": "success",
            "description": "",
            "description_tooltip": null,
            "layout": "IPY_MODEL_9448abe0d444492881e132c60f0ad15b",
            "max": 11,
            "min": 0,
            "orientation": "horizontal",
            "style": "IPY_MODEL_c9a1c923f8f746d8bfd7714df9dd7ba2",
            "value": 11
          }
        },
        "95d09829a7784ffb98b7b8012096c719": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "HTMLModel",
          "model_module_version": "1.5.0",
          "state": {
            "_dom_classes": [],
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "HTMLModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/controls",
            "_view_module_version": "1.5.0",
            "_view_name": "HTMLView",
            "description": "",
            "description_tooltip": null,
            "layout": "IPY_MODEL_489a7947f1b04f6787bd8c2b1eb19c1c",
            "placeholder": "​",
            "style": "IPY_MODEL_bd90dd854a22421a9d3e51bf078558aa",
            "value": " 11/11 [00:00&lt;00:00, 112.84it/s]"
          }
        },
        "7b96a8f4b91c4dcf92c2d470b927cda5": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "model_module_version": "1.2.0",
          "state": {
            "_model_module": "@jupyter-widgets/base",
            "_model_module_version": "1.2.0",
            "_model_name": "LayoutModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "LayoutView",
            "align_content": null,
            "align_items": null,
            "align_self": null,
            "border": null,
            "bottom": null,
            "display": null,
            "flex": null,
            "flex_flow": null,
            "grid_area": null,
            "grid_auto_columns": null,
            "grid_auto_flow": null,
            "grid_auto_rows": null,
            "grid_column": null,
            "grid_gap": null,
            "grid_row": null,
            "grid_template_areas": null,
            "grid_template_columns": null,
            "grid_template_rows": null,
            "height": null,
            "justify_content": null,
            "justify_items": null,
            "left": null,
            "margin": null,
            "max_height": null,
            "max_width": null,
            "min_height": null,
            "min_width": null,
            "object_fit": null,
            "object_position": null,
            "order": null,
            "overflow": null,
            "overflow_x": null,
            "overflow_y": null,
            "padding": null,
            "right": null,
            "top": null,
            "visibility": null,
            "width": null
          }
        },
        "180ab15da4c545be9d48a7f358c3dfe2": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "model_module_version": "1.2.0",
          "state": {
            "_model_module": "@jupyter-widgets/base",
            "_model_module_version": "1.2.0",
            "_model_name": "LayoutModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "LayoutView",
            "align_content": null,
            "align_items": null,
            "align_self": null,
            "border": null,
            "bottom": null,
            "display": null,
            "flex": null,
            "flex_flow": null,
            "grid_area": null,
            "grid_auto_columns": null,
            "grid_auto_flow": null,
            "grid_auto_rows": null,
            "grid_column": null,
            "grid_gap": null,
            "grid_row": null,
            "grid_template_areas": null,
            "grid_template_columns": null,
            "grid_template_rows": null,
            "height": null,
            "justify_content": null,
            "justify_items": null,
            "left": null,
            "margin": null,
            "max_height": null,
            "max_width": null,
            "min_height": null,
            "min_width": null,
            "object_fit": null,
            "object_position": null,
            "order": null,
            "overflow": null,
            "overflow_x": null,
            "overflow_y": null,
            "padding": null,
            "right": null,
            "top": null,
            "visibility": null,
            "width": null
          }
        },
        "92b34f04a0044697a9cdd2d617858cd8": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "DescriptionStyleModel",
          "model_module_version": "1.5.0",
          "state": {
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "DescriptionStyleModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "StyleView",
            "description_width": ""
          }
        },
        "9448abe0d444492881e132c60f0ad15b": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "model_module_version": "1.2.0",
          "state": {
            "_model_module": "@jupyter-widgets/base",
            "_model_module_version": "1.2.0",
            "_model_name": "LayoutModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "LayoutView",
            "align_content": null,
            "align_items": null,
            "align_self": null,
            "border": null,
            "bottom": null,
            "display": null,
            "flex": null,
            "flex_flow": null,
            "grid_area": null,
            "grid_auto_columns": null,
            "grid_auto_flow": null,
            "grid_auto_rows": null,
            "grid_column": null,
            "grid_gap": null,
            "grid_row": null,
            "grid_template_areas": null,
            "grid_template_columns": null,
            "grid_template_rows": null,
            "height": null,
            "justify_content": null,
            "justify_items": null,
            "left": null,
            "margin": null,
            "max_height": null,
            "max_width": null,
            "min_height": null,
            "min_width": null,
            "object_fit": null,
            "object_position": null,
            "order": null,
            "overflow": null,
            "overflow_x": null,
            "overflow_y": null,
            "padding": null,
            "right": null,
            "top": null,
            "visibility": null,
            "width": null
          }
        },
        "c9a1c923f8f746d8bfd7714df9dd7ba2": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "ProgressStyleModel",
          "model_module_version": "1.5.0",
          "state": {
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "ProgressStyleModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "StyleView",
            "bar_color": null,
            "description_width": ""
          }
        },
        "489a7947f1b04f6787bd8c2b1eb19c1c": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "model_module_version": "1.2.0",
          "state": {
            "_model_module": "@jupyter-widgets/base",
            "_model_module_version": "1.2.0",
            "_model_name": "LayoutModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "LayoutView",
            "align_content": null,
            "align_items": null,
            "align_self": null,
            "border": null,
            "bottom": null,
            "display": null,
            "flex": null,
            "flex_flow": null,
            "grid_area": null,
            "grid_auto_columns": null,
            "grid_auto_flow": null,
            "grid_auto_rows": null,
            "grid_column": null,
            "grid_gap": null,
            "grid_row": null,
            "grid_template_areas": null,
            "grid_template_columns": null,
            "grid_template_rows": null,
            "height": null,
            "justify_content": null,
            "justify_items": null,
            "left": null,
            "margin": null,
            "max_height": null,
            "max_width": null,
            "min_height": null,
            "min_width": null,
            "object_fit": null,
            "object_position": null,
            "order": null,
            "overflow": null,
            "overflow_x": null,
            "overflow_y": null,
            "padding": null,
            "right": null,
            "top": null,
            "visibility": null,
            "width": null
          }
        },
        "bd90dd854a22421a9d3e51bf078558aa": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "DescriptionStyleModel",
          "model_module_version": "1.5.0",
          "state": {
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "DescriptionStyleModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "StyleView",
            "description_width": ""
          }
        },
        "646b7becf4a04502829504d9cfedfb71": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "HBoxModel",
          "model_module_version": "1.5.0",
          "state": {
            "_dom_classes": [],
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "HBoxModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/controls",
            "_view_module_version": "1.5.0",
            "_view_name": "HBoxView",
            "box_style": "",
            "children": [
              "IPY_MODEL_4911ab5d742940259ae9c00cb66d9293",
              "IPY_MODEL_3d63ca44ccb84d0dad6824b3d33b2dd6",
              "IPY_MODEL_892ddaf2adae4ea1babdd3795899bf0d"
            ],
            "layout": "IPY_MODEL_a69da305157f497797161e07fb6f748d"
          }
        },
        "4911ab5d742940259ae9c00cb66d9293": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "HTMLModel",
          "model_module_version": "1.5.0",
          "state": {
            "_dom_classes": [],
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "HTMLModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/controls",
            "_view_module_version": "1.5.0",
            "_view_name": "HTMLView",
            "description": "",
            "description_tooltip": null,
            "layout": "IPY_MODEL_9f8b14798bf846a6a79b06fa69f8fe39",
            "placeholder": "​",
            "style": "IPY_MODEL_7ae4d744435042189fa96bc81f2cecd8",
            "value": "100%"
          }
        },
        "3d63ca44ccb84d0dad6824b3d33b2dd6": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "FloatProgressModel",
          "model_module_version": "1.5.0",
          "state": {
            "_dom_classes": [],
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "FloatProgressModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/controls",
            "_view_module_version": "1.5.0",
            "_view_name": "ProgressView",
            "bar_style": "success",
            "description": "",
            "description_tooltip": null,
            "layout": "IPY_MODEL_b86c447201de4da4b71b0c214f3b6a06",
            "max": 11,
            "min": 0,
            "orientation": "horizontal",
            "style": "IPY_MODEL_3bb772d030a74f28b857b6cf7266986e",
            "value": 11
          }
        },
        "892ddaf2adae4ea1babdd3795899bf0d": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "HTMLModel",
          "model_module_version": "1.5.0",
          "state": {
            "_dom_classes": [],
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "HTMLModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/controls",
            "_view_module_version": "1.5.0",
            "_view_name": "HTMLView",
            "description": "",
            "description_tooltip": null,
            "layout": "IPY_MODEL_401f6e163844443fa8a3135afb15c8df",
            "placeholder": "​",
            "style": "IPY_MODEL_23d158ac64a74eedad1b217f6147b4a0",
            "value": " 11/11 [00:00&lt;00:00, 916.93it/s]"
          }
        },
        "a69da305157f497797161e07fb6f748d": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "model_module_version": "1.2.0",
          "state": {
            "_model_module": "@jupyter-widgets/base",
            "_model_module_version": "1.2.0",
            "_model_name": "LayoutModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "LayoutView",
            "align_content": null,
            "align_items": null,
            "align_self": null,
            "border": null,
            "bottom": null,
            "display": null,
            "flex": null,
            "flex_flow": null,
            "grid_area": null,
            "grid_auto_columns": null,
            "grid_auto_flow": null,
            "grid_auto_rows": null,
            "grid_column": null,
            "grid_gap": null,
            "grid_row": null,
            "grid_template_areas": null,
            "grid_template_columns": null,
            "grid_template_rows": null,
            "height": null,
            "justify_content": null,
            "justify_items": null,
            "left": null,
            "margin": null,
            "max_height": null,
            "max_width": null,
            "min_height": null,
            "min_width": null,
            "object_fit": null,
            "object_position": null,
            "order": null,
            "overflow": null,
            "overflow_x": null,
            "overflow_y": null,
            "padding": null,
            "right": null,
            "top": null,
            "visibility": null,
            "width": null
          }
        },
        "9f8b14798bf846a6a79b06fa69f8fe39": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "model_module_version": "1.2.0",
          "state": {
            "_model_module": "@jupyter-widgets/base",
            "_model_module_version": "1.2.0",
            "_model_name": "LayoutModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "LayoutView",
            "align_content": null,
            "align_items": null,
            "align_self": null,
            "border": null,
            "bottom": null,
            "display": null,
            "flex": null,
            "flex_flow": null,
            "grid_area": null,
            "grid_auto_columns": null,
            "grid_auto_flow": null,
            "grid_auto_rows": null,
            "grid_column": null,
            "grid_gap": null,
            "grid_row": null,
            "grid_template_areas": null,
            "grid_template_columns": null,
            "grid_template_rows": null,
            "height": null,
            "justify_content": null,
            "justify_items": null,
            "left": null,
            "margin": null,
            "max_height": null,
            "max_width": null,
            "min_height": null,
            "min_width": null,
            "object_fit": null,
            "object_position": null,
            "order": null,
            "overflow": null,
            "overflow_x": null,
            "overflow_y": null,
            "padding": null,
            "right": null,
            "top": null,
            "visibility": null,
            "width": null
          }
        },
        "7ae4d744435042189fa96bc81f2cecd8": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "DescriptionStyleModel",
          "model_module_version": "1.5.0",
          "state": {
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "DescriptionStyleModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "StyleView",
            "description_width": ""
          }
        },
        "b86c447201de4da4b71b0c214f3b6a06": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "model_module_version": "1.2.0",
          "state": {
            "_model_module": "@jupyter-widgets/base",
            "_model_module_version": "1.2.0",
            "_model_name": "LayoutModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "LayoutView",
            "align_content": null,
            "align_items": null,
            "align_self": null,
            "border": null,
            "bottom": null,
            "display": null,
            "flex": null,
            "flex_flow": null,
            "grid_area": null,
            "grid_auto_columns": null,
            "grid_auto_flow": null,
            "grid_auto_rows": null,
            "grid_column": null,
            "grid_gap": null,
            "grid_row": null,
            "grid_template_areas": null,
            "grid_template_columns": null,
            "grid_template_rows": null,
            "height": null,
            "justify_content": null,
            "justify_items": null,
            "left": null,
            "margin": null,
            "max_height": null,
            "max_width": null,
            "min_height": null,
            "min_width": null,
            "object_fit": null,
            "object_position": null,
            "order": null,
            "overflow": null,
            "overflow_x": null,
            "overflow_y": null,
            "padding": null,
            "right": null,
            "top": null,
            "visibility": null,
            "width": null
          }
        },
        "3bb772d030a74f28b857b6cf7266986e": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "ProgressStyleModel",
          "model_module_version": "1.5.0",
          "state": {
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "ProgressStyleModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "StyleView",
            "bar_color": null,
            "description_width": ""
          }
        },
        "401f6e163844443fa8a3135afb15c8df": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "model_module_version": "1.2.0",
          "state": {
            "_model_module": "@jupyter-widgets/base",
            "_model_module_version": "1.2.0",
            "_model_name": "LayoutModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "LayoutView",
            "align_content": null,
            "align_items": null,
            "align_self": null,
            "border": null,
            "bottom": null,
            "display": null,
            "flex": null,
            "flex_flow": null,
            "grid_area": null,
            "grid_auto_columns": null,
            "grid_auto_flow": null,
            "grid_auto_rows": null,
            "grid_column": null,
            "grid_gap": null,
            "grid_row": null,
            "grid_template_areas": null,
            "grid_template_columns": null,
            "grid_template_rows": null,
            "height": null,
            "justify_content": null,
            "justify_items": null,
            "left": null,
            "margin": null,
            "max_height": null,
            "max_width": null,
            "min_height": null,
            "min_width": null,
            "object_fit": null,
            "object_position": null,
            "order": null,
            "overflow": null,
            "overflow_x": null,
            "overflow_y": null,
            "padding": null,
            "right": null,
            "top": null,
            "visibility": null,
            "width": null
          }
        },
        "23d158ac64a74eedad1b217f6147b4a0": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "DescriptionStyleModel",
          "model_module_version": "1.5.0",
          "state": {
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "DescriptionStyleModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "StyleView",
            "description_width": ""
          }
        },
        "847db1d1903d4543b1d8532629158bc6": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "HBoxModel",
          "model_module_version": "1.5.0",
          "state": {
            "_dom_classes": [],
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "HBoxModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/controls",
            "_view_module_version": "1.5.0",
            "_view_name": "HBoxView",
            "box_style": "",
            "children": [
              "IPY_MODEL_7163d2588d0b47d8809c200e34bc27f4",
              "IPY_MODEL_e1641ed76f1b45659125c3a96774bac9",
              "IPY_MODEL_976f66fb48f647a1b0941b98bb341876"
            ],
            "layout": "IPY_MODEL_23e1e281c9864a85abe5d8bbf24dc859"
          }
        },
        "7163d2588d0b47d8809c200e34bc27f4": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "HTMLModel",
          "model_module_version": "1.5.0",
          "state": {
            "_dom_classes": [],
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "HTMLModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/controls",
            "_view_module_version": "1.5.0",
            "_view_name": "HTMLView",
            "description": "",
            "description_tooltip": null,
            "layout": "IPY_MODEL_988c783606c740339c09bfe5d126f782",
            "placeholder": "​",
            "style": "IPY_MODEL_49d8c16e33af4224b30d2caccc66c140",
            "value": "100%"
          }
        },
        "e1641ed76f1b45659125c3a96774bac9": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "FloatProgressModel",
          "model_module_version": "1.5.0",
          "state": {
            "_dom_classes": [],
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "FloatProgressModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/controls",
            "_view_module_version": "1.5.0",
            "_view_name": "ProgressView",
            "bar_style": "success",
            "description": "",
            "description_tooltip": null,
            "layout": "IPY_MODEL_66a35f25189447d48ce4b3506488c35a",
            "max": 11,
            "min": 0,
            "orientation": "horizontal",
            "style": "IPY_MODEL_7a6462ef583a4d65a0e8cae6c6962bfa",
            "value": 11
          }
        },
        "976f66fb48f647a1b0941b98bb341876": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "HTMLModel",
          "model_module_version": "1.5.0",
          "state": {
            "_dom_classes": [],
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "HTMLModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/controls",
            "_view_module_version": "1.5.0",
            "_view_name": "HTMLView",
            "description": "",
            "description_tooltip": null,
            "layout": "IPY_MODEL_65cd40e423df4f69989385b92549f3dc",
            "placeholder": "​",
            "style": "IPY_MODEL_08116949b49645afb5d3a1d8232f6139",
            "value": " 11/11 [00:00&lt;00:00, 840.83it/s]"
          }
        },
        "23e1e281c9864a85abe5d8bbf24dc859": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "model_module_version": "1.2.0",
          "state": {
            "_model_module": "@jupyter-widgets/base",
            "_model_module_version": "1.2.0",
            "_model_name": "LayoutModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "LayoutView",
            "align_content": null,
            "align_items": null,
            "align_self": null,
            "border": null,
            "bottom": null,
            "display": null,
            "flex": null,
            "flex_flow": null,
            "grid_area": null,
            "grid_auto_columns": null,
            "grid_auto_flow": null,
            "grid_auto_rows": null,
            "grid_column": null,
            "grid_gap": null,
            "grid_row": null,
            "grid_template_areas": null,
            "grid_template_columns": null,
            "grid_template_rows": null,
            "height": null,
            "justify_content": null,
            "justify_items": null,
            "left": null,
            "margin": null,
            "max_height": null,
            "max_width": null,
            "min_height": null,
            "min_width": null,
            "object_fit": null,
            "object_position": null,
            "order": null,
            "overflow": null,
            "overflow_x": null,
            "overflow_y": null,
            "padding": null,
            "right": null,
            "top": null,
            "visibility": null,
            "width": null
          }
        },
        "988c783606c740339c09bfe5d126f782": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "model_module_version": "1.2.0",
          "state": {
            "_model_module": "@jupyter-widgets/base",
            "_model_module_version": "1.2.0",
            "_model_name": "LayoutModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "LayoutView",
            "align_content": null,
            "align_items": null,
            "align_self": null,
            "border": null,
            "bottom": null,
            "display": null,
            "flex": null,
            "flex_flow": null,
            "grid_area": null,
            "grid_auto_columns": null,
            "grid_auto_flow": null,
            "grid_auto_rows": null,
            "grid_column": null,
            "grid_gap": null,
            "grid_row": null,
            "grid_template_areas": null,
            "grid_template_columns": null,
            "grid_template_rows": null,
            "height": null,
            "justify_content": null,
            "justify_items": null,
            "left": null,
            "margin": null,
            "max_height": null,
            "max_width": null,
            "min_height": null,
            "min_width": null,
            "object_fit": null,
            "object_position": null,
            "order": null,
            "overflow": null,
            "overflow_x": null,
            "overflow_y": null,
            "padding": null,
            "right": null,
            "top": null,
            "visibility": null,
            "width": null
          }
        },
        "49d8c16e33af4224b30d2caccc66c140": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "DescriptionStyleModel",
          "model_module_version": "1.5.0",
          "state": {
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "DescriptionStyleModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "StyleView",
            "description_width": ""
          }
        },
        "66a35f25189447d48ce4b3506488c35a": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "model_module_version": "1.2.0",
          "state": {
            "_model_module": "@jupyter-widgets/base",
            "_model_module_version": "1.2.0",
            "_model_name": "LayoutModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "LayoutView",
            "align_content": null,
            "align_items": null,
            "align_self": null,
            "border": null,
            "bottom": null,
            "display": null,
            "flex": null,
            "flex_flow": null,
            "grid_area": null,
            "grid_auto_columns": null,
            "grid_auto_flow": null,
            "grid_auto_rows": null,
            "grid_column": null,
            "grid_gap": null,
            "grid_row": null,
            "grid_template_areas": null,
            "grid_template_columns": null,
            "grid_template_rows": null,
            "height": null,
            "justify_content": null,
            "justify_items": null,
            "left": null,
            "margin": null,
            "max_height": null,
            "max_width": null,
            "min_height": null,
            "min_width": null,
            "object_fit": null,
            "object_position": null,
            "order": null,
            "overflow": null,
            "overflow_x": null,
            "overflow_y": null,
            "padding": null,
            "right": null,
            "top": null,
            "visibility": null,
            "width": null
          }
        },
        "7a6462ef583a4d65a0e8cae6c6962bfa": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "ProgressStyleModel",
          "model_module_version": "1.5.0",
          "state": {
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "ProgressStyleModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "StyleView",
            "bar_color": null,
            "description_width": ""
          }
        },
        "65cd40e423df4f69989385b92549f3dc": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "model_module_version": "1.2.0",
          "state": {
            "_model_module": "@jupyter-widgets/base",
            "_model_module_version": "1.2.0",
            "_model_name": "LayoutModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "LayoutView",
            "align_content": null,
            "align_items": null,
            "align_self": null,
            "border": null,
            "bottom": null,
            "display": null,
            "flex": null,
            "flex_flow": null,
            "grid_area": null,
            "grid_auto_columns": null,
            "grid_auto_flow": null,
            "grid_auto_rows": null,
            "grid_column": null,
            "grid_gap": null,
            "grid_row": null,
            "grid_template_areas": null,
            "grid_template_columns": null,
            "grid_template_rows": null,
            "height": null,
            "justify_content": null,
            "justify_items": null,
            "left": null,
            "margin": null,
            "max_height": null,
            "max_width": null,
            "min_height": null,
            "min_width": null,
            "object_fit": null,
            "object_position": null,
            "order": null,
            "overflow": null,
            "overflow_x": null,
            "overflow_y": null,
            "padding": null,
            "right": null,
            "top": null,
            "visibility": null,
            "width": null
          }
        },
        "08116949b49645afb5d3a1d8232f6139": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "DescriptionStyleModel",
          "model_module_version": "1.5.0",
          "state": {
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "DescriptionStyleModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "StyleView",
            "description_width": ""
          }
        },
        "917cf9de71c446d1904037d3ff10598c": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "HBoxModel",
          "model_module_version": "1.5.0",
          "state": {
            "_dom_classes": [],
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "HBoxModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/controls",
            "_view_module_version": "1.5.0",
            "_view_name": "HBoxView",
            "box_style": "",
            "children": [
              "IPY_MODEL_a0f0732d553042e0a3be3ced90546a1e",
              "IPY_MODEL_c5bd847c659149dabfc2e1ccf99b90e4",
              "IPY_MODEL_73fe14a93678443e9b3290a07e88f6a1"
            ],
            "layout": "IPY_MODEL_4a1b367d9ccd41ca983ef892198b86b9"
          }
        },
        "a0f0732d553042e0a3be3ced90546a1e": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "HTMLModel",
          "model_module_version": "1.5.0",
          "state": {
            "_dom_classes": [],
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "HTMLModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/controls",
            "_view_module_version": "1.5.0",
            "_view_name": "HTMLView",
            "description": "",
            "description_tooltip": null,
            "layout": "IPY_MODEL_25b61ae5d8954d5a947963efae7fa41f",
            "placeholder": "​",
            "style": "IPY_MODEL_913af069fbb94c1da12f72104712121e",
            "value": "modules.json: 100%"
          }
        },
        "c5bd847c659149dabfc2e1ccf99b90e4": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "FloatProgressModel",
          "model_module_version": "1.5.0",
          "state": {
            "_dom_classes": [],
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "FloatProgressModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/controls",
            "_view_module_version": "1.5.0",
            "_view_name": "ProgressView",
            "bar_style": "success",
            "description": "",
            "description_tooltip": null,
            "layout": "IPY_MODEL_b57d5ad3287b4bb8b73a21e7038be814",
            "max": 349,
            "min": 0,
            "orientation": "horizontal",
            "style": "IPY_MODEL_5fe4516b0613424fbcbf453a7d430a2b",
            "value": 349
          }
        },
        "73fe14a93678443e9b3290a07e88f6a1": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "HTMLModel",
          "model_module_version": "1.5.0",
          "state": {
            "_dom_classes": [],
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "HTMLModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/controls",
            "_view_module_version": "1.5.0",
            "_view_name": "HTMLView",
            "description": "",
            "description_tooltip": null,
            "layout": "IPY_MODEL_2d1c40e0bfe346a9891a8583c9ef0087",
            "placeholder": "​",
            "style": "IPY_MODEL_3bc7dacd564548d181cec9088079fc73",
            "value": " 349/349 [00:00&lt;00:00, 29.8kB/s]"
          }
        },
        "4a1b367d9ccd41ca983ef892198b86b9": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "model_module_version": "1.2.0",
          "state": {
            "_model_module": "@jupyter-widgets/base",
            "_model_module_version": "1.2.0",
            "_model_name": "LayoutModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "LayoutView",
            "align_content": null,
            "align_items": null,
            "align_self": null,
            "border": null,
            "bottom": null,
            "display": null,
            "flex": null,
            "flex_flow": null,
            "grid_area": null,
            "grid_auto_columns": null,
            "grid_auto_flow": null,
            "grid_auto_rows": null,
            "grid_column": null,
            "grid_gap": null,
            "grid_row": null,
            "grid_template_areas": null,
            "grid_template_columns": null,
            "grid_template_rows": null,
            "height": null,
            "justify_content": null,
            "justify_items": null,
            "left": null,
            "margin": null,
            "max_height": null,
            "max_width": null,
            "min_height": null,
            "min_width": null,
            "object_fit": null,
            "object_position": null,
            "order": null,
            "overflow": null,
            "overflow_x": null,
            "overflow_y": null,
            "padding": null,
            "right": null,
            "top": null,
            "visibility": null,
            "width": null
          }
        },
        "25b61ae5d8954d5a947963efae7fa41f": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "model_module_version": "1.2.0",
          "state": {
            "_model_module": "@jupyter-widgets/base",
            "_model_module_version": "1.2.0",
            "_model_name": "LayoutModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "LayoutView",
            "align_content": null,
            "align_items": null,
            "align_self": null,
            "border": null,
            "bottom": null,
            "display": null,
            "flex": null,
            "flex_flow": null,
            "grid_area": null,
            "grid_auto_columns": null,
            "grid_auto_flow": null,
            "grid_auto_rows": null,
            "grid_column": null,
            "grid_gap": null,
            "grid_row": null,
            "grid_template_areas": null,
            "grid_template_columns": null,
            "grid_template_rows": null,
            "height": null,
            "justify_content": null,
            "justify_items": null,
            "left": null,
            "margin": null,
            "max_height": null,
            "max_width": null,
            "min_height": null,
            "min_width": null,
            "object_fit": null,
            "object_position": null,
            "order": null,
            "overflow": null,
            "overflow_x": null,
            "overflow_y": null,
            "padding": null,
            "right": null,
            "top": null,
            "visibility": null,
            "width": null
          }
        },
        "913af069fbb94c1da12f72104712121e": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "DescriptionStyleModel",
          "model_module_version": "1.5.0",
          "state": {
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "DescriptionStyleModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "StyleView",
            "description_width": ""
          }
        },
        "b57d5ad3287b4bb8b73a21e7038be814": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "model_module_version": "1.2.0",
          "state": {
            "_model_module": "@jupyter-widgets/base",
            "_model_module_version": "1.2.0",
            "_model_name": "LayoutModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "LayoutView",
            "align_content": null,
            "align_items": null,
            "align_self": null,
            "border": null,
            "bottom": null,
            "display": null,
            "flex": null,
            "flex_flow": null,
            "grid_area": null,
            "grid_auto_columns": null,
            "grid_auto_flow": null,
            "grid_auto_rows": null,
            "grid_column": null,
            "grid_gap": null,
            "grid_row": null,
            "grid_template_areas": null,
            "grid_template_columns": null,
            "grid_template_rows": null,
            "height": null,
            "justify_content": null,
            "justify_items": null,
            "left": null,
            "margin": null,
            "max_height": null,
            "max_width": null,
            "min_height": null,
            "min_width": null,
            "object_fit": null,
            "object_position": null,
            "order": null,
            "overflow": null,
            "overflow_x": null,
            "overflow_y": null,
            "padding": null,
            "right": null,
            "top": null,
            "visibility": null,
            "width": null
          }
        },
        "5fe4516b0613424fbcbf453a7d430a2b": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "ProgressStyleModel",
          "model_module_version": "1.5.0",
          "state": {
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "ProgressStyleModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "StyleView",
            "bar_color": null,
            "description_width": ""
          }
        },
        "2d1c40e0bfe346a9891a8583c9ef0087": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "model_module_version": "1.2.0",
          "state": {
            "_model_module": "@jupyter-widgets/base",
            "_model_module_version": "1.2.0",
            "_model_name": "LayoutModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "LayoutView",
            "align_content": null,
            "align_items": null,
            "align_self": null,
            "border": null,
            "bottom": null,
            "display": null,
            "flex": null,
            "flex_flow": null,
            "grid_area": null,
            "grid_auto_columns": null,
            "grid_auto_flow": null,
            "grid_auto_rows": null,
            "grid_column": null,
            "grid_gap": null,
            "grid_row": null,
            "grid_template_areas": null,
            "grid_template_columns": null,
            "grid_template_rows": null,
            "height": null,
            "justify_content": null,
            "justify_items": null,
            "left": null,
            "margin": null,
            "max_height": null,
            "max_width": null,
            "min_height": null,
            "min_width": null,
            "object_fit": null,
            "object_position": null,
            "order": null,
            "overflow": null,
            "overflow_x": null,
            "overflow_y": null,
            "padding": null,
            "right": null,
            "top": null,
            "visibility": null,
            "width": null
          }
        },
        "3bc7dacd564548d181cec9088079fc73": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "DescriptionStyleModel",
          "model_module_version": "1.5.0",
          "state": {
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "DescriptionStyleModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "StyleView",
            "description_width": ""
          }
        },
        "18b4d737f97945f5b56c109158d2d8cf": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "HBoxModel",
          "model_module_version": "1.5.0",
          "state": {
            "_dom_classes": [],
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "HBoxModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/controls",
            "_view_module_version": "1.5.0",
            "_view_name": "HBoxView",
            "box_style": "",
            "children": [
              "IPY_MODEL_9b715eb261574cc6ae286b3ba4a5ce5e",
              "IPY_MODEL_d88544be1bc5456691521c725a49c87f",
              "IPY_MODEL_a02ea3cc6ad645329fb54fead34c7533"
            ],
            "layout": "IPY_MODEL_c8be9055dcaa40e4b78d17362ec00632"
          }
        },
        "9b715eb261574cc6ae286b3ba4a5ce5e": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "HTMLModel",
          "model_module_version": "1.5.0",
          "state": {
            "_dom_classes": [],
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "HTMLModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/controls",
            "_view_module_version": "1.5.0",
            "_view_name": "HTMLView",
            "description": "",
            "description_tooltip": null,
            "layout": "IPY_MODEL_01fa3a36918d4a2083ebfc5830a63da0",
            "placeholder": "​",
            "style": "IPY_MODEL_0b9a20b821d84a218dcd7e7ed7d70d65",
            "value": "config_sentence_transformers.json: 100%"
          }
        },
        "d88544be1bc5456691521c725a49c87f": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "FloatProgressModel",
          "model_module_version": "1.5.0",
          "state": {
            "_dom_classes": [],
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "FloatProgressModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/controls",
            "_view_module_version": "1.5.0",
            "_view_name": "ProgressView",
            "bar_style": "success",
            "description": "",
            "description_tooltip": null,
            "layout": "IPY_MODEL_ae2fbd69796f40eb8527eb3c08258ec6",
            "max": 116,
            "min": 0,
            "orientation": "horizontal",
            "style": "IPY_MODEL_550768688aa34f06a3e0f13c1dba0e3b",
            "value": 116
          }
        },
        "a02ea3cc6ad645329fb54fead34c7533": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "HTMLModel",
          "model_module_version": "1.5.0",
          "state": {
            "_dom_classes": [],
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "HTMLModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/controls",
            "_view_module_version": "1.5.0",
            "_view_name": "HTMLView",
            "description": "",
            "description_tooltip": null,
            "layout": "IPY_MODEL_0eea9da5905149c791d9ff82eea946f1",
            "placeholder": "​",
            "style": "IPY_MODEL_41ebc1df3f4a47a7a1199078c512e129",
            "value": " 116/116 [00:00&lt;00:00, 8.82kB/s]"
          }
        },
        "c8be9055dcaa40e4b78d17362ec00632": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "model_module_version": "1.2.0",
          "state": {
            "_model_module": "@jupyter-widgets/base",
            "_model_module_version": "1.2.0",
            "_model_name": "LayoutModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "LayoutView",
            "align_content": null,
            "align_items": null,
            "align_self": null,
            "border": null,
            "bottom": null,
            "display": null,
            "flex": null,
            "flex_flow": null,
            "grid_area": null,
            "grid_auto_columns": null,
            "grid_auto_flow": null,
            "grid_auto_rows": null,
            "grid_column": null,
            "grid_gap": null,
            "grid_row": null,
            "grid_template_areas": null,
            "grid_template_columns": null,
            "grid_template_rows": null,
            "height": null,
            "justify_content": null,
            "justify_items": null,
            "left": null,
            "margin": null,
            "max_height": null,
            "max_width": null,
            "min_height": null,
            "min_width": null,
            "object_fit": null,
            "object_position": null,
            "order": null,
            "overflow": null,
            "overflow_x": null,
            "overflow_y": null,
            "padding": null,
            "right": null,
            "top": null,
            "visibility": null,
            "width": null
          }
        },
        "01fa3a36918d4a2083ebfc5830a63da0": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "model_module_version": "1.2.0",
          "state": {
            "_model_module": "@jupyter-widgets/base",
            "_model_module_version": "1.2.0",
            "_model_name": "LayoutModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "LayoutView",
            "align_content": null,
            "align_items": null,
            "align_self": null,
            "border": null,
            "bottom": null,
            "display": null,
            "flex": null,
            "flex_flow": null,
            "grid_area": null,
            "grid_auto_columns": null,
            "grid_auto_flow": null,
            "grid_auto_rows": null,
            "grid_column": null,
            "grid_gap": null,
            "grid_row": null,
            "grid_template_areas": null,
            "grid_template_columns": null,
            "grid_template_rows": null,
            "height": null,
            "justify_content": null,
            "justify_items": null,
            "left": null,
            "margin": null,
            "max_height": null,
            "max_width": null,
            "min_height": null,
            "min_width": null,
            "object_fit": null,
            "object_position": null,
            "order": null,
            "overflow": null,
            "overflow_x": null,
            "overflow_y": null,
            "padding": null,
            "right": null,
            "top": null,
            "visibility": null,
            "width": null
          }
        },
        "0b9a20b821d84a218dcd7e7ed7d70d65": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "DescriptionStyleModel",
          "model_module_version": "1.5.0",
          "state": {
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "DescriptionStyleModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "StyleView",
            "description_width": ""
          }
        },
        "ae2fbd69796f40eb8527eb3c08258ec6": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "model_module_version": "1.2.0",
          "state": {
            "_model_module": "@jupyter-widgets/base",
            "_model_module_version": "1.2.0",
            "_model_name": "LayoutModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "LayoutView",
            "align_content": null,
            "align_items": null,
            "align_self": null,
            "border": null,
            "bottom": null,
            "display": null,
            "flex": null,
            "flex_flow": null,
            "grid_area": null,
            "grid_auto_columns": null,
            "grid_auto_flow": null,
            "grid_auto_rows": null,
            "grid_column": null,
            "grid_gap": null,
            "grid_row": null,
            "grid_template_areas": null,
            "grid_template_columns": null,
            "grid_template_rows": null,
            "height": null,
            "justify_content": null,
            "justify_items": null,
            "left": null,
            "margin": null,
            "max_height": null,
            "max_width": null,
            "min_height": null,
            "min_width": null,
            "object_fit": null,
            "object_position": null,
            "order": null,
            "overflow": null,
            "overflow_x": null,
            "overflow_y": null,
            "padding": null,
            "right": null,
            "top": null,
            "visibility": null,
            "width": null
          }
        },
        "550768688aa34f06a3e0f13c1dba0e3b": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "ProgressStyleModel",
          "model_module_version": "1.5.0",
          "state": {
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "ProgressStyleModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "StyleView",
            "bar_color": null,
            "description_width": ""
          }
        },
        "0eea9da5905149c791d9ff82eea946f1": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "model_module_version": "1.2.0",
          "state": {
            "_model_module": "@jupyter-widgets/base",
            "_model_module_version": "1.2.0",
            "_model_name": "LayoutModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "LayoutView",
            "align_content": null,
            "align_items": null,
            "align_self": null,
            "border": null,
            "bottom": null,
            "display": null,
            "flex": null,
            "flex_flow": null,
            "grid_area": null,
            "grid_auto_columns": null,
            "grid_auto_flow": null,
            "grid_auto_rows": null,
            "grid_column": null,
            "grid_gap": null,
            "grid_row": null,
            "grid_template_areas": null,
            "grid_template_columns": null,
            "grid_template_rows": null,
            "height": null,
            "justify_content": null,
            "justify_items": null,
            "left": null,
            "margin": null,
            "max_height": null,
            "max_width": null,
            "min_height": null,
            "min_width": null,
            "object_fit": null,
            "object_position": null,
            "order": null,
            "overflow": null,
            "overflow_x": null,
            "overflow_y": null,
            "padding": null,
            "right": null,
            "top": null,
            "visibility": null,
            "width": null
          }
        },
        "41ebc1df3f4a47a7a1199078c512e129": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "DescriptionStyleModel",
          "model_module_version": "1.5.0",
          "state": {
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "DescriptionStyleModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "StyleView",
            "description_width": ""
          }
        },
        "c7d41b79979d4e52bebaa1439b085ee8": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "HBoxModel",
          "model_module_version": "1.5.0",
          "state": {
            "_dom_classes": [],
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "HBoxModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/controls",
            "_view_module_version": "1.5.0",
            "_view_name": "HBoxView",
            "box_style": "",
            "children": [
              "IPY_MODEL_9a6d1bfea04646f4adc753bc9e4559f8",
              "IPY_MODEL_09201e30592e448690b28882ee985c6b",
              "IPY_MODEL_1e6f7d1986a84bdcbbb621aa7bbd7305"
            ],
            "layout": "IPY_MODEL_6e5cdca32c414328847bd9aca70f273b"
          }
        },
        "9a6d1bfea04646f4adc753bc9e4559f8": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "HTMLModel",
          "model_module_version": "1.5.0",
          "state": {
            "_dom_classes": [],
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "HTMLModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/controls",
            "_view_module_version": "1.5.0",
            "_view_name": "HTMLView",
            "description": "",
            "description_tooltip": null,
            "layout": "IPY_MODEL_b588003d48d04a199546728f4cc0fde6",
            "placeholder": "​",
            "style": "IPY_MODEL_ea16b1b46032464fbcd82478ddd6a40c",
            "value": "README.md: 100%"
          }
        },
        "09201e30592e448690b28882ee985c6b": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "FloatProgressModel",
          "model_module_version": "1.5.0",
          "state": {
            "_dom_classes": [],
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "FloatProgressModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/controls",
            "_view_module_version": "1.5.0",
            "_view_name": "ProgressView",
            "bar_style": "success",
            "description": "",
            "description_tooltip": null,
            "layout": "IPY_MODEL_5f6222157d8042b4b88e9f5178653512",
            "max": 10621,
            "min": 0,
            "orientation": "horizontal",
            "style": "IPY_MODEL_219462e1e94c4b7789412ab4f2b8d50b",
            "value": 10621
          }
        },
        "1e6f7d1986a84bdcbbb621aa7bbd7305": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "HTMLModel",
          "model_module_version": "1.5.0",
          "state": {
            "_dom_classes": [],
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "HTMLModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/controls",
            "_view_module_version": "1.5.0",
            "_view_name": "HTMLView",
            "description": "",
            "description_tooltip": null,
            "layout": "IPY_MODEL_0f2c243303fd4c31befbcbf9a2703953",
            "placeholder": "​",
            "style": "IPY_MODEL_55ec1bcc7ecf4fe88a13da12c71e521b",
            "value": " 10.6k/10.6k [00:00&lt;00:00, 948kB/s]"
          }
        },
        "6e5cdca32c414328847bd9aca70f273b": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "model_module_version": "1.2.0",
          "state": {
            "_model_module": "@jupyter-widgets/base",
            "_model_module_version": "1.2.0",
            "_model_name": "LayoutModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "LayoutView",
            "align_content": null,
            "align_items": null,
            "align_self": null,
            "border": null,
            "bottom": null,
            "display": null,
            "flex": null,
            "flex_flow": null,
            "grid_area": null,
            "grid_auto_columns": null,
            "grid_auto_flow": null,
            "grid_auto_rows": null,
            "grid_column": null,
            "grid_gap": null,
            "grid_row": null,
            "grid_template_areas": null,
            "grid_template_columns": null,
            "grid_template_rows": null,
            "height": null,
            "justify_content": null,
            "justify_items": null,
            "left": null,
            "margin": null,
            "max_height": null,
            "max_width": null,
            "min_height": null,
            "min_width": null,
            "object_fit": null,
            "object_position": null,
            "order": null,
            "overflow": null,
            "overflow_x": null,
            "overflow_y": null,
            "padding": null,
            "right": null,
            "top": null,
            "visibility": null,
            "width": null
          }
        },
        "b588003d48d04a199546728f4cc0fde6": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "model_module_version": "1.2.0",
          "state": {
            "_model_module": "@jupyter-widgets/base",
            "_model_module_version": "1.2.0",
            "_model_name": "LayoutModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "LayoutView",
            "align_content": null,
            "align_items": null,
            "align_self": null,
            "border": null,
            "bottom": null,
            "display": null,
            "flex": null,
            "flex_flow": null,
            "grid_area": null,
            "grid_auto_columns": null,
            "grid_auto_flow": null,
            "grid_auto_rows": null,
            "grid_column": null,
            "grid_gap": null,
            "grid_row": null,
            "grid_template_areas": null,
            "grid_template_columns": null,
            "grid_template_rows": null,
            "height": null,
            "justify_content": null,
            "justify_items": null,
            "left": null,
            "margin": null,
            "max_height": null,
            "max_width": null,
            "min_height": null,
            "min_width": null,
            "object_fit": null,
            "object_position": null,
            "order": null,
            "overflow": null,
            "overflow_x": null,
            "overflow_y": null,
            "padding": null,
            "right": null,
            "top": null,
            "visibility": null,
            "width": null
          }
        },
        "ea16b1b46032464fbcd82478ddd6a40c": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "DescriptionStyleModel",
          "model_module_version": "1.5.0",
          "state": {
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "DescriptionStyleModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "StyleView",
            "description_width": ""
          }
        },
        "5f6222157d8042b4b88e9f5178653512": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "model_module_version": "1.2.0",
          "state": {
            "_model_module": "@jupyter-widgets/base",
            "_model_module_version": "1.2.0",
            "_model_name": "LayoutModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "LayoutView",
            "align_content": null,
            "align_items": null,
            "align_self": null,
            "border": null,
            "bottom": null,
            "display": null,
            "flex": null,
            "flex_flow": null,
            "grid_area": null,
            "grid_auto_columns": null,
            "grid_auto_flow": null,
            "grid_auto_rows": null,
            "grid_column": null,
            "grid_gap": null,
            "grid_row": null,
            "grid_template_areas": null,
            "grid_template_columns": null,
            "grid_template_rows": null,
            "height": null,
            "justify_content": null,
            "justify_items": null,
            "left": null,
            "margin": null,
            "max_height": null,
            "max_width": null,
            "min_height": null,
            "min_width": null,
            "object_fit": null,
            "object_position": null,
            "order": null,
            "overflow": null,
            "overflow_x": null,
            "overflow_y": null,
            "padding": null,
            "right": null,
            "top": null,
            "visibility": null,
            "width": null
          }
        },
        "219462e1e94c4b7789412ab4f2b8d50b": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "ProgressStyleModel",
          "model_module_version": "1.5.0",
          "state": {
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "ProgressStyleModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "StyleView",
            "bar_color": null,
            "description_width": ""
          }
        },
        "0f2c243303fd4c31befbcbf9a2703953": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "model_module_version": "1.2.0",
          "state": {
            "_model_module": "@jupyter-widgets/base",
            "_model_module_version": "1.2.0",
            "_model_name": "LayoutModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "LayoutView",
            "align_content": null,
            "align_items": null,
            "align_self": null,
            "border": null,
            "bottom": null,
            "display": null,
            "flex": null,
            "flex_flow": null,
            "grid_area": null,
            "grid_auto_columns": null,
            "grid_auto_flow": null,
            "grid_auto_rows": null,
            "grid_column": null,
            "grid_gap": null,
            "grid_row": null,
            "grid_template_areas": null,
            "grid_template_columns": null,
            "grid_template_rows": null,
            "height": null,
            "justify_content": null,
            "justify_items": null,
            "left": null,
            "margin": null,
            "max_height": null,
            "max_width": null,
            "min_height": null,
            "min_width": null,
            "object_fit": null,
            "object_position": null,
            "order": null,
            "overflow": null,
            "overflow_x": null,
            "overflow_y": null,
            "padding": null,
            "right": null,
            "top": null,
            "visibility": null,
            "width": null
          }
        },
        "55ec1bcc7ecf4fe88a13da12c71e521b": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "DescriptionStyleModel",
          "model_module_version": "1.5.0",
          "state": {
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "DescriptionStyleModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "StyleView",
            "description_width": ""
          }
        },
        "edb7e6d6a8c9418baf3cc538bc5084b1": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "HBoxModel",
          "model_module_version": "1.5.0",
          "state": {
            "_dom_classes": [],
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "HBoxModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/controls",
            "_view_module_version": "1.5.0",
            "_view_name": "HBoxView",
            "box_style": "",
            "children": [
              "IPY_MODEL_00382f4dee2744c1906bc09de6763787",
              "IPY_MODEL_5bd202d2b9e043b79a171860d3fb45d8",
              "IPY_MODEL_d178e61f3ac64c9eb3f27af5f12fce71"
            ],
            "layout": "IPY_MODEL_509a93c7ab264dcfa622cd53765cbaa7"
          }
        },
        "00382f4dee2744c1906bc09de6763787": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "HTMLModel",
          "model_module_version": "1.5.0",
          "state": {
            "_dom_classes": [],
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "HTMLModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/controls",
            "_view_module_version": "1.5.0",
            "_view_name": "HTMLView",
            "description": "",
            "description_tooltip": null,
            "layout": "IPY_MODEL_5293267b40eb4f6fb633bcd410e93b22",
            "placeholder": "​",
            "style": "IPY_MODEL_28925d888cc84557b75aac7079e50fd5",
            "value": "sentence_bert_config.json: 100%"
          }
        },
        "5bd202d2b9e043b79a171860d3fb45d8": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "FloatProgressModel",
          "model_module_version": "1.5.0",
          "state": {
            "_dom_classes": [],
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "FloatProgressModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/controls",
            "_view_module_version": "1.5.0",
            "_view_name": "ProgressView",
            "bar_style": "success",
            "description": "",
            "description_tooltip": null,
            "layout": "IPY_MODEL_114db12faa874a73b8479b256dfc124d",
            "max": 53,
            "min": 0,
            "orientation": "horizontal",
            "style": "IPY_MODEL_ec87b3217fec4aa4812bb3c1cc317f8c",
            "value": 53
          }
        },
        "d178e61f3ac64c9eb3f27af5f12fce71": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "HTMLModel",
          "model_module_version": "1.5.0",
          "state": {
            "_dom_classes": [],
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "HTMLModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/controls",
            "_view_module_version": "1.5.0",
            "_view_name": "HTMLView",
            "description": "",
            "description_tooltip": null,
            "layout": "IPY_MODEL_2f50c68c88f545f29d756abb36faee2c",
            "placeholder": "​",
            "style": "IPY_MODEL_f395239fd1c14e2fa52189b243fa1c45",
            "value": " 53.0/53.0 [00:00&lt;00:00, 4.73kB/s]"
          }
        },
        "509a93c7ab264dcfa622cd53765cbaa7": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "model_module_version": "1.2.0",
          "state": {
            "_model_module": "@jupyter-widgets/base",
            "_model_module_version": "1.2.0",
            "_model_name": "LayoutModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "LayoutView",
            "align_content": null,
            "align_items": null,
            "align_self": null,
            "border": null,
            "bottom": null,
            "display": null,
            "flex": null,
            "flex_flow": null,
            "grid_area": null,
            "grid_auto_columns": null,
            "grid_auto_flow": null,
            "grid_auto_rows": null,
            "grid_column": null,
            "grid_gap": null,
            "grid_row": null,
            "grid_template_areas": null,
            "grid_template_columns": null,
            "grid_template_rows": null,
            "height": null,
            "justify_content": null,
            "justify_items": null,
            "left": null,
            "margin": null,
            "max_height": null,
            "max_width": null,
            "min_height": null,
            "min_width": null,
            "object_fit": null,
            "object_position": null,
            "order": null,
            "overflow": null,
            "overflow_x": null,
            "overflow_y": null,
            "padding": null,
            "right": null,
            "top": null,
            "visibility": null,
            "width": null
          }
        },
        "5293267b40eb4f6fb633bcd410e93b22": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "model_module_version": "1.2.0",
          "state": {
            "_model_module": "@jupyter-widgets/base",
            "_model_module_version": "1.2.0",
            "_model_name": "LayoutModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "LayoutView",
            "align_content": null,
            "align_items": null,
            "align_self": null,
            "border": null,
            "bottom": null,
            "display": null,
            "flex": null,
            "flex_flow": null,
            "grid_area": null,
            "grid_auto_columns": null,
            "grid_auto_flow": null,
            "grid_auto_rows": null,
            "grid_column": null,
            "grid_gap": null,
            "grid_row": null,
            "grid_template_areas": null,
            "grid_template_columns": null,
            "grid_template_rows": null,
            "height": null,
            "justify_content": null,
            "justify_items": null,
            "left": null,
            "margin": null,
            "max_height": null,
            "max_width": null,
            "min_height": null,
            "min_width": null,
            "object_fit": null,
            "object_position": null,
            "order": null,
            "overflow": null,
            "overflow_x": null,
            "overflow_y": null,
            "padding": null,
            "right": null,
            "top": null,
            "visibility": null,
            "width": null
          }
        },
        "28925d888cc84557b75aac7079e50fd5": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "DescriptionStyleModel",
          "model_module_version": "1.5.0",
          "state": {
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "DescriptionStyleModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "StyleView",
            "description_width": ""
          }
        },
        "114db12faa874a73b8479b256dfc124d": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "model_module_version": "1.2.0",
          "state": {
            "_model_module": "@jupyter-widgets/base",
            "_model_module_version": "1.2.0",
            "_model_name": "LayoutModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "LayoutView",
            "align_content": null,
            "align_items": null,
            "align_self": null,
            "border": null,
            "bottom": null,
            "display": null,
            "flex": null,
            "flex_flow": null,
            "grid_area": null,
            "grid_auto_columns": null,
            "grid_auto_flow": null,
            "grid_auto_rows": null,
            "grid_column": null,
            "grid_gap": null,
            "grid_row": null,
            "grid_template_areas": null,
            "grid_template_columns": null,
            "grid_template_rows": null,
            "height": null,
            "justify_content": null,
            "justify_items": null,
            "left": null,
            "margin": null,
            "max_height": null,
            "max_width": null,
            "min_height": null,
            "min_width": null,
            "object_fit": null,
            "object_position": null,
            "order": null,
            "overflow": null,
            "overflow_x": null,
            "overflow_y": null,
            "padding": null,
            "right": null,
            "top": null,
            "visibility": null,
            "width": null
          }
        },
        "ec87b3217fec4aa4812bb3c1cc317f8c": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "ProgressStyleModel",
          "model_module_version": "1.5.0",
          "state": {
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "ProgressStyleModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "StyleView",
            "bar_color": null,
            "description_width": ""
          }
        },
        "2f50c68c88f545f29d756abb36faee2c": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "model_module_version": "1.2.0",
          "state": {
            "_model_module": "@jupyter-widgets/base",
            "_model_module_version": "1.2.0",
            "_model_name": "LayoutModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "LayoutView",
            "align_content": null,
            "align_items": null,
            "align_self": null,
            "border": null,
            "bottom": null,
            "display": null,
            "flex": null,
            "flex_flow": null,
            "grid_area": null,
            "grid_auto_columns": null,
            "grid_auto_flow": null,
            "grid_auto_rows": null,
            "grid_column": null,
            "grid_gap": null,
            "grid_row": null,
            "grid_template_areas": null,
            "grid_template_columns": null,
            "grid_template_rows": null,
            "height": null,
            "justify_content": null,
            "justify_items": null,
            "left": null,
            "margin": null,
            "max_height": null,
            "max_width": null,
            "min_height": null,
            "min_width": null,
            "object_fit": null,
            "object_position": null,
            "order": null,
            "overflow": null,
            "overflow_x": null,
            "overflow_y": null,
            "padding": null,
            "right": null,
            "top": null,
            "visibility": null,
            "width": null
          }
        },
        "f395239fd1c14e2fa52189b243fa1c45": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "DescriptionStyleModel",
          "model_module_version": "1.5.0",
          "state": {
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "DescriptionStyleModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "StyleView",
            "description_width": ""
          }
        },
        "4e66d50820314451be18f0106cf018d9": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "HBoxModel",
          "model_module_version": "1.5.0",
          "state": {
            "_dom_classes": [],
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "HBoxModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/controls",
            "_view_module_version": "1.5.0",
            "_view_name": "HBoxView",
            "box_style": "",
            "children": [
              "IPY_MODEL_6f0745faad3344ffb9a7104996bcd85b",
              "IPY_MODEL_c24bcce292a341c2b026e1a990b6c859",
              "IPY_MODEL_f33f9842b7674f5885c137e235904248"
            ],
            "layout": "IPY_MODEL_ef6a09eec3d245e292108e9c7887c648"
          }
        },
        "6f0745faad3344ffb9a7104996bcd85b": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "HTMLModel",
          "model_module_version": "1.5.0",
          "state": {
            "_dom_classes": [],
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "HTMLModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/controls",
            "_view_module_version": "1.5.0",
            "_view_name": "HTMLView",
            "description": "",
            "description_tooltip": null,
            "layout": "IPY_MODEL_04b07550df0e47aeb80a43e4d1f47d7a",
            "placeholder": "​",
            "style": "IPY_MODEL_ebd38d359a89452b9d2652f90f8d9fef",
            "value": "config.json: 100%"
          }
        },
        "c24bcce292a341c2b026e1a990b6c859": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "FloatProgressModel",
          "model_module_version": "1.5.0",
          "state": {
            "_dom_classes": [],
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "FloatProgressModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/controls",
            "_view_module_version": "1.5.0",
            "_view_name": "ProgressView",
            "bar_style": "success",
            "description": "",
            "description_tooltip": null,
            "layout": "IPY_MODEL_00c929f0c32640819be5cc03b59a92c0",
            "max": 571,
            "min": 0,
            "orientation": "horizontal",
            "style": "IPY_MODEL_e59437e8fd6142cda9fb150b1f6fb555",
            "value": 571
          }
        },
        "f33f9842b7674f5885c137e235904248": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "HTMLModel",
          "model_module_version": "1.5.0",
          "state": {
            "_dom_classes": [],
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "HTMLModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/controls",
            "_view_module_version": "1.5.0",
            "_view_name": "HTMLView",
            "description": "",
            "description_tooltip": null,
            "layout": "IPY_MODEL_873a408ae76a41bdb4f0879c551bd405",
            "placeholder": "​",
            "style": "IPY_MODEL_df61d96dbb6949bda5e9a987484b77a7",
            "value": " 571/571 [00:00&lt;00:00, 48.7kB/s]"
          }
        },
        "ef6a09eec3d245e292108e9c7887c648": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "model_module_version": "1.2.0",
          "state": {
            "_model_module": "@jupyter-widgets/base",
            "_model_module_version": "1.2.0",
            "_model_name": "LayoutModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "LayoutView",
            "align_content": null,
            "align_items": null,
            "align_self": null,
            "border": null,
            "bottom": null,
            "display": null,
            "flex": null,
            "flex_flow": null,
            "grid_area": null,
            "grid_auto_columns": null,
            "grid_auto_flow": null,
            "grid_auto_rows": null,
            "grid_column": null,
            "grid_gap": null,
            "grid_row": null,
            "grid_template_areas": null,
            "grid_template_columns": null,
            "grid_template_rows": null,
            "height": null,
            "justify_content": null,
            "justify_items": null,
            "left": null,
            "margin": null,
            "max_height": null,
            "max_width": null,
            "min_height": null,
            "min_width": null,
            "object_fit": null,
            "object_position": null,
            "order": null,
            "overflow": null,
            "overflow_x": null,
            "overflow_y": null,
            "padding": null,
            "right": null,
            "top": null,
            "visibility": null,
            "width": null
          }
        },
        "04b07550df0e47aeb80a43e4d1f47d7a": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "model_module_version": "1.2.0",
          "state": {
            "_model_module": "@jupyter-widgets/base",
            "_model_module_version": "1.2.0",
            "_model_name": "LayoutModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "LayoutView",
            "align_content": null,
            "align_items": null,
            "align_self": null,
            "border": null,
            "bottom": null,
            "display": null,
            "flex": null,
            "flex_flow": null,
            "grid_area": null,
            "grid_auto_columns": null,
            "grid_auto_flow": null,
            "grid_auto_rows": null,
            "grid_column": null,
            "grid_gap": null,
            "grid_row": null,
            "grid_template_areas": null,
            "grid_template_columns": null,
            "grid_template_rows": null,
            "height": null,
            "justify_content": null,
            "justify_items": null,
            "left": null,
            "margin": null,
            "max_height": null,
            "max_width": null,
            "min_height": null,
            "min_width": null,
            "object_fit": null,
            "object_position": null,
            "order": null,
            "overflow": null,
            "overflow_x": null,
            "overflow_y": null,
            "padding": null,
            "right": null,
            "top": null,
            "visibility": null,
            "width": null
          }
        },
        "ebd38d359a89452b9d2652f90f8d9fef": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "DescriptionStyleModel",
          "model_module_version": "1.5.0",
          "state": {
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "DescriptionStyleModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "StyleView",
            "description_width": ""
          }
        },
        "00c929f0c32640819be5cc03b59a92c0": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "model_module_version": "1.2.0",
          "state": {
            "_model_module": "@jupyter-widgets/base",
            "_model_module_version": "1.2.0",
            "_model_name": "LayoutModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "LayoutView",
            "align_content": null,
            "align_items": null,
            "align_self": null,
            "border": null,
            "bottom": null,
            "display": null,
            "flex": null,
            "flex_flow": null,
            "grid_area": null,
            "grid_auto_columns": null,
            "grid_auto_flow": null,
            "grid_auto_rows": null,
            "grid_column": null,
            "grid_gap": null,
            "grid_row": null,
            "grid_template_areas": null,
            "grid_template_columns": null,
            "grid_template_rows": null,
            "height": null,
            "justify_content": null,
            "justify_items": null,
            "left": null,
            "margin": null,
            "max_height": null,
            "max_width": null,
            "min_height": null,
            "min_width": null,
            "object_fit": null,
            "object_position": null,
            "order": null,
            "overflow": null,
            "overflow_x": null,
            "overflow_y": null,
            "padding": null,
            "right": null,
            "top": null,
            "visibility": null,
            "width": null
          }
        },
        "e59437e8fd6142cda9fb150b1f6fb555": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "ProgressStyleModel",
          "model_module_version": "1.5.0",
          "state": {
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "ProgressStyleModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "StyleView",
            "bar_color": null,
            "description_width": ""
          }
        },
        "873a408ae76a41bdb4f0879c551bd405": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "model_module_version": "1.2.0",
          "state": {
            "_model_module": "@jupyter-widgets/base",
            "_model_module_version": "1.2.0",
            "_model_name": "LayoutModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "LayoutView",
            "align_content": null,
            "align_items": null,
            "align_self": null,
            "border": null,
            "bottom": null,
            "display": null,
            "flex": null,
            "flex_flow": null,
            "grid_area": null,
            "grid_auto_columns": null,
            "grid_auto_flow": null,
            "grid_auto_rows": null,
            "grid_column": null,
            "grid_gap": null,
            "grid_row": null,
            "grid_template_areas": null,
            "grid_template_columns": null,
            "grid_template_rows": null,
            "height": null,
            "justify_content": null,
            "justify_items": null,
            "left": null,
            "margin": null,
            "max_height": null,
            "max_width": null,
            "min_height": null,
            "min_width": null,
            "object_fit": null,
            "object_position": null,
            "order": null,
            "overflow": null,
            "overflow_x": null,
            "overflow_y": null,
            "padding": null,
            "right": null,
            "top": null,
            "visibility": null,
            "width": null
          }
        },
        "df61d96dbb6949bda5e9a987484b77a7": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "DescriptionStyleModel",
          "model_module_version": "1.5.0",
          "state": {
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "DescriptionStyleModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "StyleView",
            "description_width": ""
          }
        },
        "d578666f07024c29897b773aa8c36471": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "HBoxModel",
          "model_module_version": "1.5.0",
          "state": {
            "_dom_classes": [],
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "HBoxModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/controls",
            "_view_module_version": "1.5.0",
            "_view_name": "HBoxView",
            "box_style": "",
            "children": [
              "IPY_MODEL_640ef099b7444df0ba7ddb91c462d151",
              "IPY_MODEL_662ab50b47d54419b781053f5e54b872",
              "IPY_MODEL_7026d50e81de4ccaa6558254ff62abfc"
            ],
            "layout": "IPY_MODEL_a38ced2468d5405ab47ba2611b39936f"
          }
        },
        "640ef099b7444df0ba7ddb91c462d151": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "HTMLModel",
          "model_module_version": "1.5.0",
          "state": {
            "_dom_classes": [],
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "HTMLModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/controls",
            "_view_module_version": "1.5.0",
            "_view_name": "HTMLView",
            "description": "",
            "description_tooltip": null,
            "layout": "IPY_MODEL_751ae9190e164ee7bfea2b5ffb31ca95",
            "placeholder": "​",
            "style": "IPY_MODEL_a0d8825e482e4e828c7de1e7ed0e5993",
            "value": "model.safetensors: 100%"
          }
        },
        "662ab50b47d54419b781053f5e54b872": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "FloatProgressModel",
          "model_module_version": "1.5.0",
          "state": {
            "_dom_classes": [],
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "FloatProgressModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/controls",
            "_view_module_version": "1.5.0",
            "_view_name": "ProgressView",
            "bar_style": "success",
            "description": "",
            "description_tooltip": null,
            "layout": "IPY_MODEL_bc8e982fb5a44004be1cc40594558615",
            "max": 437971872,
            "min": 0,
            "orientation": "horizontal",
            "style": "IPY_MODEL_b331e927ffce4e1fa329a4dd0b5c4a40",
            "value": 437971872
          }
        },
        "7026d50e81de4ccaa6558254ff62abfc": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "HTMLModel",
          "model_module_version": "1.5.0",
          "state": {
            "_dom_classes": [],
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "HTMLModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/controls",
            "_view_module_version": "1.5.0",
            "_view_name": "HTMLView",
            "description": "",
            "description_tooltip": null,
            "layout": "IPY_MODEL_b1554c6cba7243a4863b2faa217270e0",
            "placeholder": "​",
            "style": "IPY_MODEL_eb530f24bd874030bda74157a8e1c808",
            "value": " 438M/438M [00:01&lt;00:00, 252MB/s]"
          }
        },
        "a38ced2468d5405ab47ba2611b39936f": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "model_module_version": "1.2.0",
          "state": {
            "_model_module": "@jupyter-widgets/base",
            "_model_module_version": "1.2.0",
            "_model_name": "LayoutModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "LayoutView",
            "align_content": null,
            "align_items": null,
            "align_self": null,
            "border": null,
            "bottom": null,
            "display": null,
            "flex": null,
            "flex_flow": null,
            "grid_area": null,
            "grid_auto_columns": null,
            "grid_auto_flow": null,
            "grid_auto_rows": null,
            "grid_column": null,
            "grid_gap": null,
            "grid_row": null,
            "grid_template_areas": null,
            "grid_template_columns": null,
            "grid_template_rows": null,
            "height": null,
            "justify_content": null,
            "justify_items": null,
            "left": null,
            "margin": null,
            "max_height": null,
            "max_width": null,
            "min_height": null,
            "min_width": null,
            "object_fit": null,
            "object_position": null,
            "order": null,
            "overflow": null,
            "overflow_x": null,
            "overflow_y": null,
            "padding": null,
            "right": null,
            "top": null,
            "visibility": null,
            "width": null
          }
        },
        "751ae9190e164ee7bfea2b5ffb31ca95": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "model_module_version": "1.2.0",
          "state": {
            "_model_module": "@jupyter-widgets/base",
            "_model_module_version": "1.2.0",
            "_model_name": "LayoutModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "LayoutView",
            "align_content": null,
            "align_items": null,
            "align_self": null,
            "border": null,
            "bottom": null,
            "display": null,
            "flex": null,
            "flex_flow": null,
            "grid_area": null,
            "grid_auto_columns": null,
            "grid_auto_flow": null,
            "grid_auto_rows": null,
            "grid_column": null,
            "grid_gap": null,
            "grid_row": null,
            "grid_template_areas": null,
            "grid_template_columns": null,
            "grid_template_rows": null,
            "height": null,
            "justify_content": null,
            "justify_items": null,
            "left": null,
            "margin": null,
            "max_height": null,
            "max_width": null,
            "min_height": null,
            "min_width": null,
            "object_fit": null,
            "object_position": null,
            "order": null,
            "overflow": null,
            "overflow_x": null,
            "overflow_y": null,
            "padding": null,
            "right": null,
            "top": null,
            "visibility": null,
            "width": null
          }
        },
        "a0d8825e482e4e828c7de1e7ed0e5993": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "DescriptionStyleModel",
          "model_module_version": "1.5.0",
          "state": {
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "DescriptionStyleModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "StyleView",
            "description_width": ""
          }
        },
        "bc8e982fb5a44004be1cc40594558615": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "model_module_version": "1.2.0",
          "state": {
            "_model_module": "@jupyter-widgets/base",
            "_model_module_version": "1.2.0",
            "_model_name": "LayoutModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "LayoutView",
            "align_content": null,
            "align_items": null,
            "align_self": null,
            "border": null,
            "bottom": null,
            "display": null,
            "flex": null,
            "flex_flow": null,
            "grid_area": null,
            "grid_auto_columns": null,
            "grid_auto_flow": null,
            "grid_auto_rows": null,
            "grid_column": null,
            "grid_gap": null,
            "grid_row": null,
            "grid_template_areas": null,
            "grid_template_columns": null,
            "grid_template_rows": null,
            "height": null,
            "justify_content": null,
            "justify_items": null,
            "left": null,
            "margin": null,
            "max_height": null,
            "max_width": null,
            "min_height": null,
            "min_width": null,
            "object_fit": null,
            "object_position": null,
            "order": null,
            "overflow": null,
            "overflow_x": null,
            "overflow_y": null,
            "padding": null,
            "right": null,
            "top": null,
            "visibility": null,
            "width": null
          }
        },
        "b331e927ffce4e1fa329a4dd0b5c4a40": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "ProgressStyleModel",
          "model_module_version": "1.5.0",
          "state": {
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "ProgressStyleModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "StyleView",
            "bar_color": null,
            "description_width": ""
          }
        },
        "b1554c6cba7243a4863b2faa217270e0": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "model_module_version": "1.2.0",
          "state": {
            "_model_module": "@jupyter-widgets/base",
            "_model_module_version": "1.2.0",
            "_model_name": "LayoutModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "LayoutView",
            "align_content": null,
            "align_items": null,
            "align_self": null,
            "border": null,
            "bottom": null,
            "display": null,
            "flex": null,
            "flex_flow": null,
            "grid_area": null,
            "grid_auto_columns": null,
            "grid_auto_flow": null,
            "grid_auto_rows": null,
            "grid_column": null,
            "grid_gap": null,
            "grid_row": null,
            "grid_template_areas": null,
            "grid_template_columns": null,
            "grid_template_rows": null,
            "height": null,
            "justify_content": null,
            "justify_items": null,
            "left": null,
            "margin": null,
            "max_height": null,
            "max_width": null,
            "min_height": null,
            "min_width": null,
            "object_fit": null,
            "object_position": null,
            "order": null,
            "overflow": null,
            "overflow_x": null,
            "overflow_y": null,
            "padding": null,
            "right": null,
            "top": null,
            "visibility": null,
            "width": null
          }
        },
        "eb530f24bd874030bda74157a8e1c808": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "DescriptionStyleModel",
          "model_module_version": "1.5.0",
          "state": {
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "DescriptionStyleModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "StyleView",
            "description_width": ""
          }
        },
        "8972b726f8384c27b53d10de9161c5c3": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "HBoxModel",
          "model_module_version": "1.5.0",
          "state": {
            "_dom_classes": [],
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "HBoxModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/controls",
            "_view_module_version": "1.5.0",
            "_view_name": "HBoxView",
            "box_style": "",
            "children": [
              "IPY_MODEL_4249003994354e7dbc517d09e1ee3882",
              "IPY_MODEL_1ace14b8664c471786656c99a0bed888",
              "IPY_MODEL_0414bb5095b74f98bdafbbb57660707f"
            ],
            "layout": "IPY_MODEL_32be678b3a18428aad780c63d429c73e"
          }
        },
        "4249003994354e7dbc517d09e1ee3882": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "HTMLModel",
          "model_module_version": "1.5.0",
          "state": {
            "_dom_classes": [],
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "HTMLModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/controls",
            "_view_module_version": "1.5.0",
            "_view_name": "HTMLView",
            "description": "",
            "description_tooltip": null,
            "layout": "IPY_MODEL_c036f1444e71425c92acabb9325f7a16",
            "placeholder": "​",
            "style": "IPY_MODEL_33ab65d148fb473e81df708f11eac353",
            "value": "tokenizer_config.json: 100%"
          }
        },
        "1ace14b8664c471786656c99a0bed888": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "FloatProgressModel",
          "model_module_version": "1.5.0",
          "state": {
            "_dom_classes": [],
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "FloatProgressModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/controls",
            "_view_module_version": "1.5.0",
            "_view_name": "ProgressView",
            "bar_style": "success",
            "description": "",
            "description_tooltip": null,
            "layout": "IPY_MODEL_3f2483f387a94d40bdb37caf019da0da",
            "max": 363,
            "min": 0,
            "orientation": "horizontal",
            "style": "IPY_MODEL_5928d37584d744a986f80c239d161067",
            "value": 363
          }
        },
        "0414bb5095b74f98bdafbbb57660707f": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "HTMLModel",
          "model_module_version": "1.5.0",
          "state": {
            "_dom_classes": [],
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "HTMLModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/controls",
            "_view_module_version": "1.5.0",
            "_view_name": "HTMLView",
            "description": "",
            "description_tooltip": null,
            "layout": "IPY_MODEL_fe709112e0404cc3adc8f5e36b4dee6c",
            "placeholder": "​",
            "style": "IPY_MODEL_b19e6c67575240c4858574bd593c5b13",
            "value": " 363/363 [00:00&lt;00:00, 29.1kB/s]"
          }
        },
        "32be678b3a18428aad780c63d429c73e": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "model_module_version": "1.2.0",
          "state": {
            "_model_module": "@jupyter-widgets/base",
            "_model_module_version": "1.2.0",
            "_model_name": "LayoutModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "LayoutView",
            "align_content": null,
            "align_items": null,
            "align_self": null,
            "border": null,
            "bottom": null,
            "display": null,
            "flex": null,
            "flex_flow": null,
            "grid_area": null,
            "grid_auto_columns": null,
            "grid_auto_flow": null,
            "grid_auto_rows": null,
            "grid_column": null,
            "grid_gap": null,
            "grid_row": null,
            "grid_template_areas": null,
            "grid_template_columns": null,
            "grid_template_rows": null,
            "height": null,
            "justify_content": null,
            "justify_items": null,
            "left": null,
            "margin": null,
            "max_height": null,
            "max_width": null,
            "min_height": null,
            "min_width": null,
            "object_fit": null,
            "object_position": null,
            "order": null,
            "overflow": null,
            "overflow_x": null,
            "overflow_y": null,
            "padding": null,
            "right": null,
            "top": null,
            "visibility": null,
            "width": null
          }
        },
        "c036f1444e71425c92acabb9325f7a16": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "model_module_version": "1.2.0",
          "state": {
            "_model_module": "@jupyter-widgets/base",
            "_model_module_version": "1.2.0",
            "_model_name": "LayoutModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "LayoutView",
            "align_content": null,
            "align_items": null,
            "align_self": null,
            "border": null,
            "bottom": null,
            "display": null,
            "flex": null,
            "flex_flow": null,
            "grid_area": null,
            "grid_auto_columns": null,
            "grid_auto_flow": null,
            "grid_auto_rows": null,
            "grid_column": null,
            "grid_gap": null,
            "grid_row": null,
            "grid_template_areas": null,
            "grid_template_columns": null,
            "grid_template_rows": null,
            "height": null,
            "justify_content": null,
            "justify_items": null,
            "left": null,
            "margin": null,
            "max_height": null,
            "max_width": null,
            "min_height": null,
            "min_width": null,
            "object_fit": null,
            "object_position": null,
            "order": null,
            "overflow": null,
            "overflow_x": null,
            "overflow_y": null,
            "padding": null,
            "right": null,
            "top": null,
            "visibility": null,
            "width": null
          }
        },
        "33ab65d148fb473e81df708f11eac353": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "DescriptionStyleModel",
          "model_module_version": "1.5.0",
          "state": {
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "DescriptionStyleModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "StyleView",
            "description_width": ""
          }
        },
        "3f2483f387a94d40bdb37caf019da0da": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "model_module_version": "1.2.0",
          "state": {
            "_model_module": "@jupyter-widgets/base",
            "_model_module_version": "1.2.0",
            "_model_name": "LayoutModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "LayoutView",
            "align_content": null,
            "align_items": null,
            "align_self": null,
            "border": null,
            "bottom": null,
            "display": null,
            "flex": null,
            "flex_flow": null,
            "grid_area": null,
            "grid_auto_columns": null,
            "grid_auto_flow": null,
            "grid_auto_rows": null,
            "grid_column": null,
            "grid_gap": null,
            "grid_row": null,
            "grid_template_areas": null,
            "grid_template_columns": null,
            "grid_template_rows": null,
            "height": null,
            "justify_content": null,
            "justify_items": null,
            "left": null,
            "margin": null,
            "max_height": null,
            "max_width": null,
            "min_height": null,
            "min_width": null,
            "object_fit": null,
            "object_position": null,
            "order": null,
            "overflow": null,
            "overflow_x": null,
            "overflow_y": null,
            "padding": null,
            "right": null,
            "top": null,
            "visibility": null,
            "width": null
          }
        },
        "5928d37584d744a986f80c239d161067": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "ProgressStyleModel",
          "model_module_version": "1.5.0",
          "state": {
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "ProgressStyleModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "StyleView",
            "bar_color": null,
            "description_width": ""
          }
        },
        "fe709112e0404cc3adc8f5e36b4dee6c": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "model_module_version": "1.2.0",
          "state": {
            "_model_module": "@jupyter-widgets/base",
            "_model_module_version": "1.2.0",
            "_model_name": "LayoutModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "LayoutView",
            "align_content": null,
            "align_items": null,
            "align_self": null,
            "border": null,
            "bottom": null,
            "display": null,
            "flex": null,
            "flex_flow": null,
            "grid_area": null,
            "grid_auto_columns": null,
            "grid_auto_flow": null,
            "grid_auto_rows": null,
            "grid_column": null,
            "grid_gap": null,
            "grid_row": null,
            "grid_template_areas": null,
            "grid_template_columns": null,
            "grid_template_rows": null,
            "height": null,
            "justify_content": null,
            "justify_items": null,
            "left": null,
            "margin": null,
            "max_height": null,
            "max_width": null,
            "min_height": null,
            "min_width": null,
            "object_fit": null,
            "object_position": null,
            "order": null,
            "overflow": null,
            "overflow_x": null,
            "overflow_y": null,
            "padding": null,
            "right": null,
            "top": null,
            "visibility": null,
            "width": null
          }
        },
        "b19e6c67575240c4858574bd593c5b13": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "DescriptionStyleModel",
          "model_module_version": "1.5.0",
          "state": {
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "DescriptionStyleModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "StyleView",
            "description_width": ""
          }
        },
        "561f4a0b88a4483ebe99a0272bfe542e": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "HBoxModel",
          "model_module_version": "1.5.0",
          "state": {
            "_dom_classes": [],
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "HBoxModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/controls",
            "_view_module_version": "1.5.0",
            "_view_name": "HBoxView",
            "box_style": "",
            "children": [
              "IPY_MODEL_eec84441708343518e35fee9bc4be234",
              "IPY_MODEL_7caa2e0a1ffc44df81804b743615829c",
              "IPY_MODEL_b7a9da9514b941f38b95f959caba67a3"
            ],
            "layout": "IPY_MODEL_716a5878c7684e23a8ade614b7029313"
          }
        },
        "eec84441708343518e35fee9bc4be234": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "HTMLModel",
          "model_module_version": "1.5.0",
          "state": {
            "_dom_classes": [],
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "HTMLModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/controls",
            "_view_module_version": "1.5.0",
            "_view_name": "HTMLView",
            "description": "",
            "description_tooltip": null,
            "layout": "IPY_MODEL_bff0f96b6765448c8a0fc62b0c7daab7",
            "placeholder": "​",
            "style": "IPY_MODEL_c8d8373fa5eb4ed59d7833901775d179",
            "value": "vocab.txt: 100%"
          }
        },
        "7caa2e0a1ffc44df81804b743615829c": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "FloatProgressModel",
          "model_module_version": "1.5.0",
          "state": {
            "_dom_classes": [],
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "FloatProgressModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/controls",
            "_view_module_version": "1.5.0",
            "_view_name": "ProgressView",
            "bar_style": "success",
            "description": "",
            "description_tooltip": null,
            "layout": "IPY_MODEL_8afe34f0fb7a43a6a4357141a20c3297",
            "max": 231536,
            "min": 0,
            "orientation": "horizontal",
            "style": "IPY_MODEL_c93112f1a32840f5a63604bbdfb060c5",
            "value": 231536
          }
        },
        "b7a9da9514b941f38b95f959caba67a3": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "HTMLModel",
          "model_module_version": "1.5.0",
          "state": {
            "_dom_classes": [],
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "HTMLModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/controls",
            "_view_module_version": "1.5.0",
            "_view_name": "HTMLView",
            "description": "",
            "description_tooltip": null,
            "layout": "IPY_MODEL_342f30219a9f47e4ad3035a8a50a54b0",
            "placeholder": "​",
            "style": "IPY_MODEL_3d5c8db5bfbe434b85699b68ca7538ab",
            "value": " 232k/232k [00:00&lt;00:00, 547kB/s]"
          }
        },
        "716a5878c7684e23a8ade614b7029313": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "model_module_version": "1.2.0",
          "state": {
            "_model_module": "@jupyter-widgets/base",
            "_model_module_version": "1.2.0",
            "_model_name": "LayoutModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "LayoutView",
            "align_content": null,
            "align_items": null,
            "align_self": null,
            "border": null,
            "bottom": null,
            "display": null,
            "flex": null,
            "flex_flow": null,
            "grid_area": null,
            "grid_auto_columns": null,
            "grid_auto_flow": null,
            "grid_auto_rows": null,
            "grid_column": null,
            "grid_gap": null,
            "grid_row": null,
            "grid_template_areas": null,
            "grid_template_columns": null,
            "grid_template_rows": null,
            "height": null,
            "justify_content": null,
            "justify_items": null,
            "left": null,
            "margin": null,
            "max_height": null,
            "max_width": null,
            "min_height": null,
            "min_width": null,
            "object_fit": null,
            "object_position": null,
            "order": null,
            "overflow": null,
            "overflow_x": null,
            "overflow_y": null,
            "padding": null,
            "right": null,
            "top": null,
            "visibility": null,
            "width": null
          }
        },
        "bff0f96b6765448c8a0fc62b0c7daab7": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "model_module_version": "1.2.0",
          "state": {
            "_model_module": "@jupyter-widgets/base",
            "_model_module_version": "1.2.0",
            "_model_name": "LayoutModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "LayoutView",
            "align_content": null,
            "align_items": null,
            "align_self": null,
            "border": null,
            "bottom": null,
            "display": null,
            "flex": null,
            "flex_flow": null,
            "grid_area": null,
            "grid_auto_columns": null,
            "grid_auto_flow": null,
            "grid_auto_rows": null,
            "grid_column": null,
            "grid_gap": null,
            "grid_row": null,
            "grid_template_areas": null,
            "grid_template_columns": null,
            "grid_template_rows": null,
            "height": null,
            "justify_content": null,
            "justify_items": null,
            "left": null,
            "margin": null,
            "max_height": null,
            "max_width": null,
            "min_height": null,
            "min_width": null,
            "object_fit": null,
            "object_position": null,
            "order": null,
            "overflow": null,
            "overflow_x": null,
            "overflow_y": null,
            "padding": null,
            "right": null,
            "top": null,
            "visibility": null,
            "width": null
          }
        },
        "c8d8373fa5eb4ed59d7833901775d179": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "DescriptionStyleModel",
          "model_module_version": "1.5.0",
          "state": {
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "DescriptionStyleModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "StyleView",
            "description_width": ""
          }
        },
        "8afe34f0fb7a43a6a4357141a20c3297": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "model_module_version": "1.2.0",
          "state": {
            "_model_module": "@jupyter-widgets/base",
            "_model_module_version": "1.2.0",
            "_model_name": "LayoutModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "LayoutView",
            "align_content": null,
            "align_items": null,
            "align_self": null,
            "border": null,
            "bottom": null,
            "display": null,
            "flex": null,
            "flex_flow": null,
            "grid_area": null,
            "grid_auto_columns": null,
            "grid_auto_flow": null,
            "grid_auto_rows": null,
            "grid_column": null,
            "grid_gap": null,
            "grid_row": null,
            "grid_template_areas": null,
            "grid_template_columns": null,
            "grid_template_rows": null,
            "height": null,
            "justify_content": null,
            "justify_items": null,
            "left": null,
            "margin": null,
            "max_height": null,
            "max_width": null,
            "min_height": null,
            "min_width": null,
            "object_fit": null,
            "object_position": null,
            "order": null,
            "overflow": null,
            "overflow_x": null,
            "overflow_y": null,
            "padding": null,
            "right": null,
            "top": null,
            "visibility": null,
            "width": null
          }
        },
        "c93112f1a32840f5a63604bbdfb060c5": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "ProgressStyleModel",
          "model_module_version": "1.5.0",
          "state": {
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "ProgressStyleModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "StyleView",
            "bar_color": null,
            "description_width": ""
          }
        },
        "342f30219a9f47e4ad3035a8a50a54b0": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "model_module_version": "1.2.0",
          "state": {
            "_model_module": "@jupyter-widgets/base",
            "_model_module_version": "1.2.0",
            "_model_name": "LayoutModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "LayoutView",
            "align_content": null,
            "align_items": null,
            "align_self": null,
            "border": null,
            "bottom": null,
            "display": null,
            "flex": null,
            "flex_flow": null,
            "grid_area": null,
            "grid_auto_columns": null,
            "grid_auto_flow": null,
            "grid_auto_rows": null,
            "grid_column": null,
            "grid_gap": null,
            "grid_row": null,
            "grid_template_areas": null,
            "grid_template_columns": null,
            "grid_template_rows": null,
            "height": null,
            "justify_content": null,
            "justify_items": null,
            "left": null,
            "margin": null,
            "max_height": null,
            "max_width": null,
            "min_height": null,
            "min_width": null,
            "object_fit": null,
            "object_position": null,
            "order": null,
            "overflow": null,
            "overflow_x": null,
            "overflow_y": null,
            "padding": null,
            "right": null,
            "top": null,
            "visibility": null,
            "width": null
          }
        },
        "3d5c8db5bfbe434b85699b68ca7538ab": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "DescriptionStyleModel",
          "model_module_version": "1.5.0",
          "state": {
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "DescriptionStyleModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "StyleView",
            "description_width": ""
          }
        },
        "311e46ee70694985af3a43ddd29b5d48": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "HBoxModel",
          "model_module_version": "1.5.0",
          "state": {
            "_dom_classes": [],
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "HBoxModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/controls",
            "_view_module_version": "1.5.0",
            "_view_name": "HBoxView",
            "box_style": "",
            "children": [
              "IPY_MODEL_d7f60361121344e58ab84284553713b3",
              "IPY_MODEL_ea37dd51374449e799db32052166736e",
              "IPY_MODEL_46bd841bfe734478a4013d12784bb95c"
            ],
            "layout": "IPY_MODEL_2211b11bc7314d279dbec814880195eb"
          }
        },
        "d7f60361121344e58ab84284553713b3": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "HTMLModel",
          "model_module_version": "1.5.0",
          "state": {
            "_dom_classes": [],
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "HTMLModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/controls",
            "_view_module_version": "1.5.0",
            "_view_name": "HTMLView",
            "description": "",
            "description_tooltip": null,
            "layout": "IPY_MODEL_6ef3dfc72fac41019ab3daf266d4605f",
            "placeholder": "​",
            "style": "IPY_MODEL_964ec1d142a343a6924c805d9fc5b6d1",
            "value": "tokenizer.json: 100%"
          }
        },
        "ea37dd51374449e799db32052166736e": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "FloatProgressModel",
          "model_module_version": "1.5.0",
          "state": {
            "_dom_classes": [],
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "FloatProgressModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/controls",
            "_view_module_version": "1.5.0",
            "_view_name": "ProgressView",
            "bar_style": "success",
            "description": "",
            "description_tooltip": null,
            "layout": "IPY_MODEL_2222c590e54a4bb6be7c0f08dfc833f1",
            "max": 466021,
            "min": 0,
            "orientation": "horizontal",
            "style": "IPY_MODEL_debc092429e445898728741e66dbae28",
            "value": 466021
          }
        },
        "46bd841bfe734478a4013d12784bb95c": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "HTMLModel",
          "model_module_version": "1.5.0",
          "state": {
            "_dom_classes": [],
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "HTMLModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/controls",
            "_view_module_version": "1.5.0",
            "_view_name": "HTMLView",
            "description": "",
            "description_tooltip": null,
            "layout": "IPY_MODEL_403ff93e2f6c41b084337485ef83cfc2",
            "placeholder": "​",
            "style": "IPY_MODEL_0fb9563c385d40fc9853a72b743121d4",
            "value": " 466k/466k [00:00&lt;00:00, 627kB/s]"
          }
        },
        "2211b11bc7314d279dbec814880195eb": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "model_module_version": "1.2.0",
          "state": {
            "_model_module": "@jupyter-widgets/base",
            "_model_module_version": "1.2.0",
            "_model_name": "LayoutModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "LayoutView",
            "align_content": null,
            "align_items": null,
            "align_self": null,
            "border": null,
            "bottom": null,
            "display": null,
            "flex": null,
            "flex_flow": null,
            "grid_area": null,
            "grid_auto_columns": null,
            "grid_auto_flow": null,
            "grid_auto_rows": null,
            "grid_column": null,
            "grid_gap": null,
            "grid_row": null,
            "grid_template_areas": null,
            "grid_template_columns": null,
            "grid_template_rows": null,
            "height": null,
            "justify_content": null,
            "justify_items": null,
            "left": null,
            "margin": null,
            "max_height": null,
            "max_width": null,
            "min_height": null,
            "min_width": null,
            "object_fit": null,
            "object_position": null,
            "order": null,
            "overflow": null,
            "overflow_x": null,
            "overflow_y": null,
            "padding": null,
            "right": null,
            "top": null,
            "visibility": null,
            "width": null
          }
        },
        "6ef3dfc72fac41019ab3daf266d4605f": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "model_module_version": "1.2.0",
          "state": {
            "_model_module": "@jupyter-widgets/base",
            "_model_module_version": "1.2.0",
            "_model_name": "LayoutModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "LayoutView",
            "align_content": null,
            "align_items": null,
            "align_self": null,
            "border": null,
            "bottom": null,
            "display": null,
            "flex": null,
            "flex_flow": null,
            "grid_area": null,
            "grid_auto_columns": null,
            "grid_auto_flow": null,
            "grid_auto_rows": null,
            "grid_column": null,
            "grid_gap": null,
            "grid_row": null,
            "grid_template_areas": null,
            "grid_template_columns": null,
            "grid_template_rows": null,
            "height": null,
            "justify_content": null,
            "justify_items": null,
            "left": null,
            "margin": null,
            "max_height": null,
            "max_width": null,
            "min_height": null,
            "min_width": null,
            "object_fit": null,
            "object_position": null,
            "order": null,
            "overflow": null,
            "overflow_x": null,
            "overflow_y": null,
            "padding": null,
            "right": null,
            "top": null,
            "visibility": null,
            "width": null
          }
        },
        "964ec1d142a343a6924c805d9fc5b6d1": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "DescriptionStyleModel",
          "model_module_version": "1.5.0",
          "state": {
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "DescriptionStyleModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "StyleView",
            "description_width": ""
          }
        },
        "2222c590e54a4bb6be7c0f08dfc833f1": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "model_module_version": "1.2.0",
          "state": {
            "_model_module": "@jupyter-widgets/base",
            "_model_module_version": "1.2.0",
            "_model_name": "LayoutModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "LayoutView",
            "align_content": null,
            "align_items": null,
            "align_self": null,
            "border": null,
            "bottom": null,
            "display": null,
            "flex": null,
            "flex_flow": null,
            "grid_area": null,
            "grid_auto_columns": null,
            "grid_auto_flow": null,
            "grid_auto_rows": null,
            "grid_column": null,
            "grid_gap": null,
            "grid_row": null,
            "grid_template_areas": null,
            "grid_template_columns": null,
            "grid_template_rows": null,
            "height": null,
            "justify_content": null,
            "justify_items": null,
            "left": null,
            "margin": null,
            "max_height": null,
            "max_width": null,
            "min_height": null,
            "min_width": null,
            "object_fit": null,
            "object_position": null,
            "order": null,
            "overflow": null,
            "overflow_x": null,
            "overflow_y": null,
            "padding": null,
            "right": null,
            "top": null,
            "visibility": null,
            "width": null
          }
        },
        "debc092429e445898728741e66dbae28": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "ProgressStyleModel",
          "model_module_version": "1.5.0",
          "state": {
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "ProgressStyleModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "StyleView",
            "bar_color": null,
            "description_width": ""
          }
        },
        "403ff93e2f6c41b084337485ef83cfc2": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "model_module_version": "1.2.0",
          "state": {
            "_model_module": "@jupyter-widgets/base",
            "_model_module_version": "1.2.0",
            "_model_name": "LayoutModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "LayoutView",
            "align_content": null,
            "align_items": null,
            "align_self": null,
            "border": null,
            "bottom": null,
            "display": null,
            "flex": null,
            "flex_flow": null,
            "grid_area": null,
            "grid_auto_columns": null,
            "grid_auto_flow": null,
            "grid_auto_rows": null,
            "grid_column": null,
            "grid_gap": null,
            "grid_row": null,
            "grid_template_areas": null,
            "grid_template_columns": null,
            "grid_template_rows": null,
            "height": null,
            "justify_content": null,
            "justify_items": null,
            "left": null,
            "margin": null,
            "max_height": null,
            "max_width": null,
            "min_height": null,
            "min_width": null,
            "object_fit": null,
            "object_position": null,
            "order": null,
            "overflow": null,
            "overflow_x": null,
            "overflow_y": null,
            "padding": null,
            "right": null,
            "top": null,
            "visibility": null,
            "width": null
          }
        },
        "0fb9563c385d40fc9853a72b743121d4": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "DescriptionStyleModel",
          "model_module_version": "1.5.0",
          "state": {
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "DescriptionStyleModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "StyleView",
            "description_width": ""
          }
        },
        "93b1d3bc25b14cc6832ef660b0208579": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "HBoxModel",
          "model_module_version": "1.5.0",
          "state": {
            "_dom_classes": [],
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "HBoxModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/controls",
            "_view_module_version": "1.5.0",
            "_view_name": "HBoxView",
            "box_style": "",
            "children": [
              "IPY_MODEL_2d695f86703840059af0f07f22abe0e5",
              "IPY_MODEL_b381d2fb48cb4c7da9770f77a8e25dc6",
              "IPY_MODEL_70416e4ca3e442d7a35c28dc326259c3"
            ],
            "layout": "IPY_MODEL_2b7c301c186f475bb7ba43df3ac32d1e"
          }
        },
        "2d695f86703840059af0f07f22abe0e5": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "HTMLModel",
          "model_module_version": "1.5.0",
          "state": {
            "_dom_classes": [],
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "HTMLModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/controls",
            "_view_module_version": "1.5.0",
            "_view_name": "HTMLView",
            "description": "",
            "description_tooltip": null,
            "layout": "IPY_MODEL_7a2905afb99741ffac956a2a6cbb52cd",
            "placeholder": "​",
            "style": "IPY_MODEL_4bc1838d2bb8439cbe610a3da0b32f01",
            "value": "special_tokens_map.json: 100%"
          }
        },
        "b381d2fb48cb4c7da9770f77a8e25dc6": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "FloatProgressModel",
          "model_module_version": "1.5.0",
          "state": {
            "_dom_classes": [],
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "FloatProgressModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/controls",
            "_view_module_version": "1.5.0",
            "_view_name": "ProgressView",
            "bar_style": "success",
            "description": "",
            "description_tooltip": null,
            "layout": "IPY_MODEL_e389e0e85bf242ca9f687c8f7e47a9ec",
            "max": 239,
            "min": 0,
            "orientation": "horizontal",
            "style": "IPY_MODEL_974021bd7cf945f08b115f501ce085ee",
            "value": 239
          }
        },
        "70416e4ca3e442d7a35c28dc326259c3": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "HTMLModel",
          "model_module_version": "1.5.0",
          "state": {
            "_dom_classes": [],
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "HTMLModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/controls",
            "_view_module_version": "1.5.0",
            "_view_name": "HTMLView",
            "description": "",
            "description_tooltip": null,
            "layout": "IPY_MODEL_c1b195650c8b47179bee7dc5337bac2e",
            "placeholder": "​",
            "style": "IPY_MODEL_7c3849ffcdce40a0baa395d3ea102f03",
            "value": " 239/239 [00:00&lt;00:00, 21.4kB/s]"
          }
        },
        "2b7c301c186f475bb7ba43df3ac32d1e": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "model_module_version": "1.2.0",
          "state": {
            "_model_module": "@jupyter-widgets/base",
            "_model_module_version": "1.2.0",
            "_model_name": "LayoutModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "LayoutView",
            "align_content": null,
            "align_items": null,
            "align_self": null,
            "border": null,
            "bottom": null,
            "display": null,
            "flex": null,
            "flex_flow": null,
            "grid_area": null,
            "grid_auto_columns": null,
            "grid_auto_flow": null,
            "grid_auto_rows": null,
            "grid_column": null,
            "grid_gap": null,
            "grid_row": null,
            "grid_template_areas": null,
            "grid_template_columns": null,
            "grid_template_rows": null,
            "height": null,
            "justify_content": null,
            "justify_items": null,
            "left": null,
            "margin": null,
            "max_height": null,
            "max_width": null,
            "min_height": null,
            "min_width": null,
            "object_fit": null,
            "object_position": null,
            "order": null,
            "overflow": null,
            "overflow_x": null,
            "overflow_y": null,
            "padding": null,
            "right": null,
            "top": null,
            "visibility": null,
            "width": null
          }
        },
        "7a2905afb99741ffac956a2a6cbb52cd": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "model_module_version": "1.2.0",
          "state": {
            "_model_module": "@jupyter-widgets/base",
            "_model_module_version": "1.2.0",
            "_model_name": "LayoutModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "LayoutView",
            "align_content": null,
            "align_items": null,
            "align_self": null,
            "border": null,
            "bottom": null,
            "display": null,
            "flex": null,
            "flex_flow": null,
            "grid_area": null,
            "grid_auto_columns": null,
            "grid_auto_flow": null,
            "grid_auto_rows": null,
            "grid_column": null,
            "grid_gap": null,
            "grid_row": null,
            "grid_template_areas": null,
            "grid_template_columns": null,
            "grid_template_rows": null,
            "height": null,
            "justify_content": null,
            "justify_items": null,
            "left": null,
            "margin": null,
            "max_height": null,
            "max_width": null,
            "min_height": null,
            "min_width": null,
            "object_fit": null,
            "object_position": null,
            "order": null,
            "overflow": null,
            "overflow_x": null,
            "overflow_y": null,
            "padding": null,
            "right": null,
            "top": null,
            "visibility": null,
            "width": null
          }
        },
        "4bc1838d2bb8439cbe610a3da0b32f01": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "DescriptionStyleModel",
          "model_module_version": "1.5.0",
          "state": {
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "DescriptionStyleModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "StyleView",
            "description_width": ""
          }
        },
        "e389e0e85bf242ca9f687c8f7e47a9ec": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "model_module_version": "1.2.0",
          "state": {
            "_model_module": "@jupyter-widgets/base",
            "_model_module_version": "1.2.0",
            "_model_name": "LayoutModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "LayoutView",
            "align_content": null,
            "align_items": null,
            "align_self": null,
            "border": null,
            "bottom": null,
            "display": null,
            "flex": null,
            "flex_flow": null,
            "grid_area": null,
            "grid_auto_columns": null,
            "grid_auto_flow": null,
            "grid_auto_rows": null,
            "grid_column": null,
            "grid_gap": null,
            "grid_row": null,
            "grid_template_areas": null,
            "grid_template_columns": null,
            "grid_template_rows": null,
            "height": null,
            "justify_content": null,
            "justify_items": null,
            "left": null,
            "margin": null,
            "max_height": null,
            "max_width": null,
            "min_height": null,
            "min_width": null,
            "object_fit": null,
            "object_position": null,
            "order": null,
            "overflow": null,
            "overflow_x": null,
            "overflow_y": null,
            "padding": null,
            "right": null,
            "top": null,
            "visibility": null,
            "width": null
          }
        },
        "974021bd7cf945f08b115f501ce085ee": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "ProgressStyleModel",
          "model_module_version": "1.5.0",
          "state": {
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "ProgressStyleModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "StyleView",
            "bar_color": null,
            "description_width": ""
          }
        },
        "c1b195650c8b47179bee7dc5337bac2e": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "model_module_version": "1.2.0",
          "state": {
            "_model_module": "@jupyter-widgets/base",
            "_model_module_version": "1.2.0",
            "_model_name": "LayoutModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "LayoutView",
            "align_content": null,
            "align_items": null,
            "align_self": null,
            "border": null,
            "bottom": null,
            "display": null,
            "flex": null,
            "flex_flow": null,
            "grid_area": null,
            "grid_auto_columns": null,
            "grid_auto_flow": null,
            "grid_auto_rows": null,
            "grid_column": null,
            "grid_gap": null,
            "grid_row": null,
            "grid_template_areas": null,
            "grid_template_columns": null,
            "grid_template_rows": null,
            "height": null,
            "justify_content": null,
            "justify_items": null,
            "left": null,
            "margin": null,
            "max_height": null,
            "max_width": null,
            "min_height": null,
            "min_width": null,
            "object_fit": null,
            "object_position": null,
            "order": null,
            "overflow": null,
            "overflow_x": null,
            "overflow_y": null,
            "padding": null,
            "right": null,
            "top": null,
            "visibility": null,
            "width": null
          }
        },
        "7c3849ffcdce40a0baa395d3ea102f03": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "DescriptionStyleModel",
          "model_module_version": "1.5.0",
          "state": {
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "DescriptionStyleModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "StyleView",
            "description_width": ""
          }
        },
        "4fa081728b234291bc569b2df0611996": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "HBoxModel",
          "model_module_version": "1.5.0",
          "state": {
            "_dom_classes": [],
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "HBoxModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/controls",
            "_view_module_version": "1.5.0",
            "_view_name": "HBoxView",
            "box_style": "",
            "children": [
              "IPY_MODEL_5cd36aef0a70430d808d413ec3e47c8b",
              "IPY_MODEL_7870b145821b40dc9e9a839748dca05a",
              "IPY_MODEL_efbb30ec5b864c029eb7ddb61389d1b4"
            ],
            "layout": "IPY_MODEL_21e9be74d6e145efaf45e58cb5709f32"
          }
        },
        "5cd36aef0a70430d808d413ec3e47c8b": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "HTMLModel",
          "model_module_version": "1.5.0",
          "state": {
            "_dom_classes": [],
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "HTMLModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/controls",
            "_view_module_version": "1.5.0",
            "_view_name": "HTMLView",
            "description": "",
            "description_tooltip": null,
            "layout": "IPY_MODEL_ef48aa6295584710a2a13a6a8fbb276c",
            "placeholder": "​",
            "style": "IPY_MODEL_60740ca54036463197c5d039f51298d2",
            "value": "1_Pooling/config.json: 100%"
          }
        },
        "7870b145821b40dc9e9a839748dca05a": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "FloatProgressModel",
          "model_module_version": "1.5.0",
          "state": {
            "_dom_classes": [],
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "FloatProgressModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/controls",
            "_view_module_version": "1.5.0",
            "_view_name": "ProgressView",
            "bar_style": "success",
            "description": "",
            "description_tooltip": null,
            "layout": "IPY_MODEL_d8dca4a7ce8c4f5aa2b0785ab41705db",
            "max": 190,
            "min": 0,
            "orientation": "horizontal",
            "style": "IPY_MODEL_29ca0ba7f9114016a164ce51196ef061",
            "value": 190
          }
        },
        "efbb30ec5b864c029eb7ddb61389d1b4": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "HTMLModel",
          "model_module_version": "1.5.0",
          "state": {
            "_dom_classes": [],
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "HTMLModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/controls",
            "_view_module_version": "1.5.0",
            "_view_name": "HTMLView",
            "description": "",
            "description_tooltip": null,
            "layout": "IPY_MODEL_5513f837752649fea50bd9c0ea209660",
            "placeholder": "​",
            "style": "IPY_MODEL_c5061f4db4fd4c3790df4d71d2e1ab4b",
            "value": " 190/190 [00:00&lt;00:00, 14.4kB/s]"
          }
        },
        "21e9be74d6e145efaf45e58cb5709f32": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "model_module_version": "1.2.0",
          "state": {
            "_model_module": "@jupyter-widgets/base",
            "_model_module_version": "1.2.0",
            "_model_name": "LayoutModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "LayoutView",
            "align_content": null,
            "align_items": null,
            "align_self": null,
            "border": null,
            "bottom": null,
            "display": null,
            "flex": null,
            "flex_flow": null,
            "grid_area": null,
            "grid_auto_columns": null,
            "grid_auto_flow": null,
            "grid_auto_rows": null,
            "grid_column": null,
            "grid_gap": null,
            "grid_row": null,
            "grid_template_areas": null,
            "grid_template_columns": null,
            "grid_template_rows": null,
            "height": null,
            "justify_content": null,
            "justify_items": null,
            "left": null,
            "margin": null,
            "max_height": null,
            "max_width": null,
            "min_height": null,
            "min_width": null,
            "object_fit": null,
            "object_position": null,
            "order": null,
            "overflow": null,
            "overflow_x": null,
            "overflow_y": null,
            "padding": null,
            "right": null,
            "top": null,
            "visibility": null,
            "width": null
          }
        },
        "ef48aa6295584710a2a13a6a8fbb276c": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "model_module_version": "1.2.0",
          "state": {
            "_model_module": "@jupyter-widgets/base",
            "_model_module_version": "1.2.0",
            "_model_name": "LayoutModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "LayoutView",
            "align_content": null,
            "align_items": null,
            "align_self": null,
            "border": null,
            "bottom": null,
            "display": null,
            "flex": null,
            "flex_flow": null,
            "grid_area": null,
            "grid_auto_columns": null,
            "grid_auto_flow": null,
            "grid_auto_rows": null,
            "grid_column": null,
            "grid_gap": null,
            "grid_row": null,
            "grid_template_areas": null,
            "grid_template_columns": null,
            "grid_template_rows": null,
            "height": null,
            "justify_content": null,
            "justify_items": null,
            "left": null,
            "margin": null,
            "max_height": null,
            "max_width": null,
            "min_height": null,
            "min_width": null,
            "object_fit": null,
            "object_position": null,
            "order": null,
            "overflow": null,
            "overflow_x": null,
            "overflow_y": null,
            "padding": null,
            "right": null,
            "top": null,
            "visibility": null,
            "width": null
          }
        },
        "60740ca54036463197c5d039f51298d2": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "DescriptionStyleModel",
          "model_module_version": "1.5.0",
          "state": {
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "DescriptionStyleModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "StyleView",
            "description_width": ""
          }
        },
        "d8dca4a7ce8c4f5aa2b0785ab41705db": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "model_module_version": "1.2.0",
          "state": {
            "_model_module": "@jupyter-widgets/base",
            "_model_module_version": "1.2.0",
            "_model_name": "LayoutModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "LayoutView",
            "align_content": null,
            "align_items": null,
            "align_self": null,
            "border": null,
            "bottom": null,
            "display": null,
            "flex": null,
            "flex_flow": null,
            "grid_area": null,
            "grid_auto_columns": null,
            "grid_auto_flow": null,
            "grid_auto_rows": null,
            "grid_column": null,
            "grid_gap": null,
            "grid_row": null,
            "grid_template_areas": null,
            "grid_template_columns": null,
            "grid_template_rows": null,
            "height": null,
            "justify_content": null,
            "justify_items": null,
            "left": null,
            "margin": null,
            "max_height": null,
            "max_width": null,
            "min_height": null,
            "min_width": null,
            "object_fit": null,
            "object_position": null,
            "order": null,
            "overflow": null,
            "overflow_x": null,
            "overflow_y": null,
            "padding": null,
            "right": null,
            "top": null,
            "visibility": null,
            "width": null
          }
        },
        "29ca0ba7f9114016a164ce51196ef061": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "ProgressStyleModel",
          "model_module_version": "1.5.0",
          "state": {
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "ProgressStyleModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "StyleView",
            "bar_color": null,
            "description_width": ""
          }
        },
        "5513f837752649fea50bd9c0ea209660": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "model_module_version": "1.2.0",
          "state": {
            "_model_module": "@jupyter-widgets/base",
            "_model_module_version": "1.2.0",
            "_model_name": "LayoutModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "LayoutView",
            "align_content": null,
            "align_items": null,
            "align_self": null,
            "border": null,
            "bottom": null,
            "display": null,
            "flex": null,
            "flex_flow": null,
            "grid_area": null,
            "grid_auto_columns": null,
            "grid_auto_flow": null,
            "grid_auto_rows": null,
            "grid_column": null,
            "grid_gap": null,
            "grid_row": null,
            "grid_template_areas": null,
            "grid_template_columns": null,
            "grid_template_rows": null,
            "height": null,
            "justify_content": null,
            "justify_items": null,
            "left": null,
            "margin": null,
            "max_height": null,
            "max_width": null,
            "min_height": null,
            "min_width": null,
            "object_fit": null,
            "object_position": null,
            "order": null,
            "overflow": null,
            "overflow_x": null,
            "overflow_y": null,
            "padding": null,
            "right": null,
            "top": null,
            "visibility": null,
            "width": null
          }
        },
        "c5061f4db4fd4c3790df4d71d2e1ab4b": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "DescriptionStyleModel",
          "model_module_version": "1.5.0",
          "state": {
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "DescriptionStyleModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "StyleView",
            "description_width": ""
          }
        },
        "66aa4dd8cfe040ad96d92733e55ad1c5": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "HBoxModel",
          "model_module_version": "1.5.0",
          "state": {
            "_dom_classes": [],
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "HBoxModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/controls",
            "_view_module_version": "1.5.0",
            "_view_name": "HBoxView",
            "box_style": "",
            "children": [
              "IPY_MODEL_0250334a34dd41a38a7135ca6da6c234",
              "IPY_MODEL_ce073b1d172543b4aab5950ebfdf1459",
              "IPY_MODEL_4dec3d51a88e483f8bdb84d57c30bf70"
            ],
            "layout": "IPY_MODEL_aa002c60e0a44244af89f6c13ffb718f"
          }
        },
        "0250334a34dd41a38a7135ca6da6c234": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "HTMLModel",
          "model_module_version": "1.5.0",
          "state": {
            "_dom_classes": [],
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "HTMLModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/controls",
            "_view_module_version": "1.5.0",
            "_view_name": "HTMLView",
            "description": "",
            "description_tooltip": null,
            "layout": "IPY_MODEL_6f573805b7ee49aca5d6e06670b585e6",
            "placeholder": "​",
            "style": "IPY_MODEL_17e995d86e68426b93147a7f1430700c",
            "value": "100%"
          }
        },
        "ce073b1d172543b4aab5950ebfdf1459": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "FloatProgressModel",
          "model_module_version": "1.5.0",
          "state": {
            "_dom_classes": [],
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "FloatProgressModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/controls",
            "_view_module_version": "1.5.0",
            "_view_name": "ProgressView",
            "bar_style": "success",
            "description": "",
            "description_tooltip": null,
            "layout": "IPY_MODEL_5cb0d781db454b59bf5ca34cea5fb991",
            "max": 35,
            "min": 0,
            "orientation": "horizontal",
            "style": "IPY_MODEL_57b442944ac346758c164c037d80dcd5",
            "value": 35
          }
        },
        "4dec3d51a88e483f8bdb84d57c30bf70": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "HTMLModel",
          "model_module_version": "1.5.0",
          "state": {
            "_dom_classes": [],
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "HTMLModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/controls",
            "_view_module_version": "1.5.0",
            "_view_name": "HTMLView",
            "description": "",
            "description_tooltip": null,
            "layout": "IPY_MODEL_670637917949438082271350428bbd5c",
            "placeholder": "​",
            "style": "IPY_MODEL_4ed404f08720460f965e5bf773e4919d",
            "value": " 35/35 [00:02&lt;00:00, 30.61it/s]"
          }
        },
        "aa002c60e0a44244af89f6c13ffb718f": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "model_module_version": "1.2.0",
          "state": {
            "_model_module": "@jupyter-widgets/base",
            "_model_module_version": "1.2.0",
            "_model_name": "LayoutModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "LayoutView",
            "align_content": null,
            "align_items": null,
            "align_self": null,
            "border": null,
            "bottom": null,
            "display": null,
            "flex": null,
            "flex_flow": null,
            "grid_area": null,
            "grid_auto_columns": null,
            "grid_auto_flow": null,
            "grid_auto_rows": null,
            "grid_column": null,
            "grid_gap": null,
            "grid_row": null,
            "grid_template_areas": null,
            "grid_template_columns": null,
            "grid_template_rows": null,
            "height": null,
            "justify_content": null,
            "justify_items": null,
            "left": null,
            "margin": null,
            "max_height": null,
            "max_width": null,
            "min_height": null,
            "min_width": null,
            "object_fit": null,
            "object_position": null,
            "order": null,
            "overflow": null,
            "overflow_x": null,
            "overflow_y": null,
            "padding": null,
            "right": null,
            "top": null,
            "visibility": null,
            "width": null
          }
        },
        "6f573805b7ee49aca5d6e06670b585e6": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "model_module_version": "1.2.0",
          "state": {
            "_model_module": "@jupyter-widgets/base",
            "_model_module_version": "1.2.0",
            "_model_name": "LayoutModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "LayoutView",
            "align_content": null,
            "align_items": null,
            "align_self": null,
            "border": null,
            "bottom": null,
            "display": null,
            "flex": null,
            "flex_flow": null,
            "grid_area": null,
            "grid_auto_columns": null,
            "grid_auto_flow": null,
            "grid_auto_rows": null,
            "grid_column": null,
            "grid_gap": null,
            "grid_row": null,
            "grid_template_areas": null,
            "grid_template_columns": null,
            "grid_template_rows": null,
            "height": null,
            "justify_content": null,
            "justify_items": null,
            "left": null,
            "margin": null,
            "max_height": null,
            "max_width": null,
            "min_height": null,
            "min_width": null,
            "object_fit": null,
            "object_position": null,
            "order": null,
            "overflow": null,
            "overflow_x": null,
            "overflow_y": null,
            "padding": null,
            "right": null,
            "top": null,
            "visibility": null,
            "width": null
          }
        },
        "17e995d86e68426b93147a7f1430700c": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "DescriptionStyleModel",
          "model_module_version": "1.5.0",
          "state": {
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "DescriptionStyleModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "StyleView",
            "description_width": ""
          }
        },
        "5cb0d781db454b59bf5ca34cea5fb991": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "model_module_version": "1.2.0",
          "state": {
            "_model_module": "@jupyter-widgets/base",
            "_model_module_version": "1.2.0",
            "_model_name": "LayoutModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "LayoutView",
            "align_content": null,
            "align_items": null,
            "align_self": null,
            "border": null,
            "bottom": null,
            "display": null,
            "flex": null,
            "flex_flow": null,
            "grid_area": null,
            "grid_auto_columns": null,
            "grid_auto_flow": null,
            "grid_auto_rows": null,
            "grid_column": null,
            "grid_gap": null,
            "grid_row": null,
            "grid_template_areas": null,
            "grid_template_columns": null,
            "grid_template_rows": null,
            "height": null,
            "justify_content": null,
            "justify_items": null,
            "left": null,
            "margin": null,
            "max_height": null,
            "max_width": null,
            "min_height": null,
            "min_width": null,
            "object_fit": null,
            "object_position": null,
            "order": null,
            "overflow": null,
            "overflow_x": null,
            "overflow_y": null,
            "padding": null,
            "right": null,
            "top": null,
            "visibility": null,
            "width": null
          }
        },
        "57b442944ac346758c164c037d80dcd5": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "ProgressStyleModel",
          "model_module_version": "1.5.0",
          "state": {
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "ProgressStyleModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "StyleView",
            "bar_color": null,
            "description_width": ""
          }
        },
        "670637917949438082271350428bbd5c": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "model_module_version": "1.2.0",
          "state": {
            "_model_module": "@jupyter-widgets/base",
            "_model_module_version": "1.2.0",
            "_model_name": "LayoutModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "LayoutView",
            "align_content": null,
            "align_items": null,
            "align_self": null,
            "border": null,
            "bottom": null,
            "display": null,
            "flex": null,
            "flex_flow": null,
            "grid_area": null,
            "grid_auto_columns": null,
            "grid_auto_flow": null,
            "grid_auto_rows": null,
            "grid_column": null,
            "grid_gap": null,
            "grid_row": null,
            "grid_template_areas": null,
            "grid_template_columns": null,
            "grid_template_rows": null,
            "height": null,
            "justify_content": null,
            "justify_items": null,
            "left": null,
            "margin": null,
            "max_height": null,
            "max_width": null,
            "min_height": null,
            "min_width": null,
            "object_fit": null,
            "object_position": null,
            "order": null,
            "overflow": null,
            "overflow_x": null,
            "overflow_y": null,
            "padding": null,
            "right": null,
            "top": null,
            "visibility": null,
            "width": null
          }
        },
        "4ed404f08720460f965e5bf773e4919d": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "DescriptionStyleModel",
          "model_module_version": "1.5.0",
          "state": {
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "DescriptionStyleModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "StyleView",
            "description_width": ""
          }
        }
      }
    }
  },
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "view-in-github",
        "colab_type": "text"
      },
      "source": [
        "<a href=\"https://colab.research.google.com/github/mhmd2015/AI/blob/main/RAG_Test1.ipynb\" target=\"_parent\"><img src=\"https://colab.research.google.com/assets/colab-badge.svg\" alt=\"Open In Colab\"/></a>"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 1,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "nosAub-JdLaJ",
        "outputId": "949a9c0c-d11f-4e9a-8228-7b22eb0e1324"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "[INFO] Running in Google Colab, installing requirements.\n",
            "Requirement already satisfied: torch in /usr/local/lib/python3.10/dist-packages (2.5.1+cu121)\n",
            "Requirement already satisfied: filelock in /usr/local/lib/python3.10/dist-packages (from torch) (3.16.1)\n",
            "Requirement already satisfied: typing-extensions>=4.8.0 in /usr/local/lib/python3.10/dist-packages (from torch) (4.12.2)\n",
            "Requirement already satisfied: networkx in /usr/local/lib/python3.10/dist-packages (from torch) (3.4.2)\n",
            "Requirement already satisfied: jinja2 in /usr/local/lib/python3.10/dist-packages (from torch) (3.1.4)\n",
            "Requirement already satisfied: fsspec in /usr/local/lib/python3.10/dist-packages (from torch) (2024.10.0)\n",
            "Requirement already satisfied: sympy==1.13.1 in /usr/local/lib/python3.10/dist-packages (from torch) (1.13.1)\n",
            "Requirement already satisfied: mpmath<1.4,>=1.1.0 in /usr/local/lib/python3.10/dist-packages (from sympy==1.13.1->torch) (1.3.0)\n",
            "Requirement already satisfied: MarkupSafe>=2.0 in /usr/local/lib/python3.10/dist-packages (from jinja2->torch) (3.0.2)\n",
            "Collecting PyMuPDF\n",
            "  Downloading PyMuPDF-1.24.14-cp39-abi3-manylinux2014_x86_64.manylinux_2_17_x86_64.whl.metadata (3.4 kB)\n",
            "Downloading PyMuPDF-1.24.14-cp39-abi3-manylinux2014_x86_64.manylinux_2_17_x86_64.whl (19.8 MB)\n",
            "\u001b[2K   \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m19.8/19.8 MB\u001b[0m \u001b[31m84.6 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
            "\u001b[?25hInstalling collected packages: PyMuPDF\n",
            "Successfully installed PyMuPDF-1.24.14\n",
            "Requirement already satisfied: tqdm in /usr/local/lib/python3.10/dist-packages (4.66.6)\n",
            "Requirement already satisfied: sentence-transformers in /usr/local/lib/python3.10/dist-packages (3.2.1)\n",
            "Requirement already satisfied: transformers<5.0.0,>=4.41.0 in /usr/local/lib/python3.10/dist-packages (from sentence-transformers) (4.46.2)\n",
            "Requirement already satisfied: tqdm in /usr/local/lib/python3.10/dist-packages (from sentence-transformers) (4.66.6)\n",
            "Requirement already satisfied: torch>=1.11.0 in /usr/local/lib/python3.10/dist-packages (from sentence-transformers) (2.5.1+cu121)\n",
            "Requirement already satisfied: scikit-learn in /usr/local/lib/python3.10/dist-packages (from sentence-transformers) (1.5.2)\n",
            "Requirement already satisfied: scipy in /usr/local/lib/python3.10/dist-packages (from sentence-transformers) (1.13.1)\n",
            "Requirement already satisfied: huggingface-hub>=0.20.0 in /usr/local/lib/python3.10/dist-packages (from sentence-transformers) (0.26.2)\n",
            "Requirement already satisfied: Pillow in /usr/local/lib/python3.10/dist-packages (from sentence-transformers) (11.0.0)\n",
            "Requirement already satisfied: filelock in /usr/local/lib/python3.10/dist-packages (from huggingface-hub>=0.20.0->sentence-transformers) (3.16.1)\n",
            "Requirement already satisfied: fsspec>=2023.5.0 in /usr/local/lib/python3.10/dist-packages (from huggingface-hub>=0.20.0->sentence-transformers) (2024.10.0)\n",
            "Requirement already satisfied: packaging>=20.9 in /usr/local/lib/python3.10/dist-packages (from huggingface-hub>=0.20.0->sentence-transformers) (24.2)\n",
            "Requirement already satisfied: pyyaml>=5.1 in /usr/local/lib/python3.10/dist-packages (from huggingface-hub>=0.20.0->sentence-transformers) (6.0.2)\n",
            "Requirement already satisfied: requests in /usr/local/lib/python3.10/dist-packages (from huggingface-hub>=0.20.0->sentence-transformers) (2.32.3)\n",
            "Requirement already satisfied: typing-extensions>=3.7.4.3 in /usr/local/lib/python3.10/dist-packages (from huggingface-hub>=0.20.0->sentence-transformers) (4.12.2)\n",
            "Requirement already satisfied: networkx in /usr/local/lib/python3.10/dist-packages (from torch>=1.11.0->sentence-transformers) (3.4.2)\n",
            "Requirement already satisfied: jinja2 in /usr/local/lib/python3.10/dist-packages (from torch>=1.11.0->sentence-transformers) (3.1.4)\n",
            "Requirement already satisfied: sympy==1.13.1 in /usr/local/lib/python3.10/dist-packages (from torch>=1.11.0->sentence-transformers) (1.13.1)\n",
            "Requirement already satisfied: mpmath<1.4,>=1.1.0 in /usr/local/lib/python3.10/dist-packages (from sympy==1.13.1->torch>=1.11.0->sentence-transformers) (1.3.0)\n",
            "Requirement already satisfied: numpy>=1.17 in /usr/local/lib/python3.10/dist-packages (from transformers<5.0.0,>=4.41.0->sentence-transformers) (1.26.4)\n",
            "Requirement already satisfied: regex!=2019.12.17 in /usr/local/lib/python3.10/dist-packages (from transformers<5.0.0,>=4.41.0->sentence-transformers) (2024.9.11)\n",
            "Requirement already satisfied: safetensors>=0.4.1 in /usr/local/lib/python3.10/dist-packages (from transformers<5.0.0,>=4.41.0->sentence-transformers) (0.4.5)\n",
            "Requirement already satisfied: tokenizers<0.21,>=0.20 in /usr/local/lib/python3.10/dist-packages (from transformers<5.0.0,>=4.41.0->sentence-transformers) (0.20.3)\n",
            "Requirement already satisfied: joblib>=1.2.0 in /usr/local/lib/python3.10/dist-packages (from scikit-learn->sentence-transformers) (1.4.2)\n",
            "Requirement already satisfied: threadpoolctl>=3.1.0 in /usr/local/lib/python3.10/dist-packages (from scikit-learn->sentence-transformers) (3.5.0)\n",
            "Requirement already satisfied: MarkupSafe>=2.0 in /usr/local/lib/python3.10/dist-packages (from jinja2->torch>=1.11.0->sentence-transformers) (3.0.2)\n",
            "Requirement already satisfied: charset-normalizer<4,>=2 in /usr/local/lib/python3.10/dist-packages (from requests->huggingface-hub>=0.20.0->sentence-transformers) (3.4.0)\n",
            "Requirement already satisfied: idna<4,>=2.5 in /usr/local/lib/python3.10/dist-packages (from requests->huggingface-hub>=0.20.0->sentence-transformers) (3.10)\n",
            "Requirement already satisfied: urllib3<3,>=1.21.1 in /usr/local/lib/python3.10/dist-packages (from requests->huggingface-hub>=0.20.0->sentence-transformers) (2.2.3)\n",
            "Requirement already satisfied: certifi>=2017.4.17 in /usr/local/lib/python3.10/dist-packages (from requests->huggingface-hub>=0.20.0->sentence-transformers) (2024.8.30)\n",
            "Requirement already satisfied: accelerate in /usr/local/lib/python3.10/dist-packages (1.1.1)\n",
            "Requirement already satisfied: huggingface-hub>=0.21.0 in /usr/local/lib/python3.10/dist-packages (from accelerate) (0.26.2)\n",
            "Requirement already satisfied: numpy<3.0.0,>=1.17 in /usr/local/lib/python3.10/dist-packages (from accelerate) (1.26.4)\n",
            "Requirement already satisfied: packaging>=20.0 in /usr/local/lib/python3.10/dist-packages (from accelerate) (24.2)\n",
            "Requirement already satisfied: psutil in /usr/local/lib/python3.10/dist-packages (from accelerate) (5.9.5)\n",
            "Requirement already satisfied: pyyaml in /usr/local/lib/python3.10/dist-packages (from accelerate) (6.0.2)\n",
            "Requirement already satisfied: safetensors>=0.4.3 in /usr/local/lib/python3.10/dist-packages (from accelerate) (0.4.5)\n",
            "Requirement already satisfied: torch>=1.10.0 in /usr/local/lib/python3.10/dist-packages (from accelerate) (2.5.1+cu121)\n",
            "Requirement already satisfied: filelock in /usr/local/lib/python3.10/dist-packages (from huggingface-hub>=0.21.0->accelerate) (3.16.1)\n",
            "Requirement already satisfied: fsspec>=2023.5.0 in /usr/local/lib/python3.10/dist-packages (from huggingface-hub>=0.21.0->accelerate) (2024.10.0)\n",
            "Requirement already satisfied: requests in /usr/local/lib/python3.10/dist-packages (from huggingface-hub>=0.21.0->accelerate) (2.32.3)\n",
            "Requirement already satisfied: tqdm>=4.42.1 in /usr/local/lib/python3.10/dist-packages (from huggingface-hub>=0.21.0->accelerate) (4.66.6)\n",
            "Requirement already satisfied: typing-extensions>=3.7.4.3 in /usr/local/lib/python3.10/dist-packages (from huggingface-hub>=0.21.0->accelerate) (4.12.2)\n",
            "Requirement already satisfied: networkx in /usr/local/lib/python3.10/dist-packages (from torch>=1.10.0->accelerate) (3.4.2)\n",
            "Requirement already satisfied: jinja2 in /usr/local/lib/python3.10/dist-packages (from torch>=1.10.0->accelerate) (3.1.4)\n",
            "Requirement already satisfied: sympy==1.13.1 in /usr/local/lib/python3.10/dist-packages (from torch>=1.10.0->accelerate) (1.13.1)\n",
            "Requirement already satisfied: mpmath<1.4,>=1.1.0 in /usr/local/lib/python3.10/dist-packages (from sympy==1.13.1->torch>=1.10.0->accelerate) (1.3.0)\n",
            "Requirement already satisfied: MarkupSafe>=2.0 in /usr/local/lib/python3.10/dist-packages (from jinja2->torch>=1.10.0->accelerate) (3.0.2)\n",
            "Requirement already satisfied: charset-normalizer<4,>=2 in /usr/local/lib/python3.10/dist-packages (from requests->huggingface-hub>=0.21.0->accelerate) (3.4.0)\n",
            "Requirement already satisfied: idna<4,>=2.5 in /usr/local/lib/python3.10/dist-packages (from requests->huggingface-hub>=0.21.0->accelerate) (3.10)\n",
            "Requirement already satisfied: urllib3<3,>=1.21.1 in /usr/local/lib/python3.10/dist-packages (from requests->huggingface-hub>=0.21.0->accelerate) (2.2.3)\n",
            "Requirement already satisfied: certifi>=2017.4.17 in /usr/local/lib/python3.10/dist-packages (from requests->huggingface-hub>=0.21.0->accelerate) (2024.8.30)\n",
            "Collecting bitsandbytes\n",
            "  Downloading bitsandbytes-0.44.1-py3-none-manylinux_2_24_x86_64.whl.metadata (3.5 kB)\n",
            "Requirement already satisfied: torch in /usr/local/lib/python3.10/dist-packages (from bitsandbytes) (2.5.1+cu121)\n",
            "Requirement already satisfied: numpy in /usr/local/lib/python3.10/dist-packages (from bitsandbytes) (1.26.4)\n",
            "Requirement already satisfied: filelock in /usr/local/lib/python3.10/dist-packages (from torch->bitsandbytes) (3.16.1)\n",
            "Requirement already satisfied: typing-extensions>=4.8.0 in /usr/local/lib/python3.10/dist-packages (from torch->bitsandbytes) (4.12.2)\n",
            "Requirement already satisfied: networkx in /usr/local/lib/python3.10/dist-packages (from torch->bitsandbytes) (3.4.2)\n",
            "Requirement already satisfied: jinja2 in /usr/local/lib/python3.10/dist-packages (from torch->bitsandbytes) (3.1.4)\n",
            "Requirement already satisfied: fsspec in /usr/local/lib/python3.10/dist-packages (from torch->bitsandbytes) (2024.10.0)\n",
            "Requirement already satisfied: sympy==1.13.1 in /usr/local/lib/python3.10/dist-packages (from torch->bitsandbytes) (1.13.1)\n",
            "Requirement already satisfied: mpmath<1.4,>=1.1.0 in /usr/local/lib/python3.10/dist-packages (from sympy==1.13.1->torch->bitsandbytes) (1.3.0)\n",
            "Requirement already satisfied: MarkupSafe>=2.0 in /usr/local/lib/python3.10/dist-packages (from jinja2->torch->bitsandbytes) (3.0.2)\n",
            "Downloading bitsandbytes-0.44.1-py3-none-manylinux_2_24_x86_64.whl (122.4 MB)\n",
            "\u001b[2K   \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m122.4/122.4 MB\u001b[0m \u001b[31m18.2 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
            "\u001b[?25hInstalling collected packages: bitsandbytes\n",
            "Successfully installed bitsandbytes-0.44.1\n",
            "Collecting flash-attn\n",
            "  Downloading flash_attn-2.7.0.post2.tar.gz (2.7 MB)\n",
            "\u001b[2K     \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m2.7/2.7 MB\u001b[0m \u001b[31m70.0 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
            "\u001b[?25h  Preparing metadata (setup.py) ... \u001b[?25l\u001b[?25hdone\n",
            "Requirement already satisfied: torch in /usr/local/lib/python3.10/dist-packages (from flash-attn) (2.5.1+cu121)\n",
            "Requirement already satisfied: einops in /usr/local/lib/python3.10/dist-packages (from flash-attn) (0.8.0)\n",
            "Requirement already satisfied: filelock in /usr/local/lib/python3.10/dist-packages (from torch->flash-attn) (3.16.1)\n",
            "Requirement already satisfied: typing-extensions>=4.8.0 in /usr/local/lib/python3.10/dist-packages (from torch->flash-attn) (4.12.2)\n",
            "Requirement already satisfied: networkx in /usr/local/lib/python3.10/dist-packages (from torch->flash-attn) (3.4.2)\n",
            "Requirement already satisfied: jinja2 in /usr/local/lib/python3.10/dist-packages (from torch->flash-attn) (3.1.4)\n",
            "Requirement already satisfied: fsspec in /usr/local/lib/python3.10/dist-packages (from torch->flash-attn) (2024.10.0)\n",
            "Requirement already satisfied: sympy==1.13.1 in /usr/local/lib/python3.10/dist-packages (from torch->flash-attn) (1.13.1)\n",
            "Requirement already satisfied: mpmath<1.4,>=1.1.0 in /usr/local/lib/python3.10/dist-packages (from sympy==1.13.1->torch->flash-attn) (1.3.0)\n",
            "Requirement already satisfied: MarkupSafe>=2.0 in /usr/local/lib/python3.10/dist-packages (from jinja2->torch->flash-attn) (3.0.2)\n",
            "Building wheels for collected packages: flash-attn\n",
            "  Building wheel for flash-attn (setup.py) ... \u001b[?25l\u001b[?25hdone\n",
            "  Created wheel for flash-attn: filename=flash_attn-2.7.0.post2-cp310-cp310-linux_x86_64.whl size=183291101 sha256=16a849d51b95cf8e47a6e6cd36826e9ffbbc068a8546e7e3501a598bd70905a6\n",
            "  Stored in directory: /root/.cache/pip/wheels/bf/e3/ed/5e845387d52f2debd1bafb847bf3d774d3f0a3c8e31b1dc948\n",
            "Successfully built flash-attn\n",
            "Installing collected packages: flash-attn\n",
            "Successfully installed flash-attn-2.7.0.post2\n"
          ]
        }
      ],
      "source": [
        "# Perform Google Colab installs (if running in Google Colab)\n",
        "import os\n",
        "\n",
        "if \"COLAB_GPU\" in os.environ:\n",
        "    print(\"[INFO] Running in Google Colab, installing requirements.\")\n",
        "    !pip install -U torch # requires torch 2.1.1+ (for efficient sdpa implementation)\n",
        "    !pip install PyMuPDF # for reading PDFs with Python\n",
        "    !pip install tqdm # for progress bars\n",
        "    !pip install sentence-transformers # for embedding models\n",
        "    !pip install accelerate # for quantization model loading\n",
        "    !pip install bitsandbytes # for quantizing models (less storage space)\n",
        "    !pip install flash-attn --no-build-isolation # for faster attention mechanism = faster LLM inference"
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# Download PDF file\n",
        "import os\n",
        "import requests\n",
        "\n",
        "# Get PDF document\n",
        "pdf_path = \"2205.07690v1.pdf\"\n",
        "\n",
        "# Download PDF if it doesn't already exist\n",
        "if not os.path.exists(pdf_path):\n",
        "  print(\"File doesn't exist, downloading...\")\n",
        "\n",
        "  # The URL of the PDF you want to download\n",
        "  url = \"https://arxiv.org/pdf/2205.07690v1\"\n",
        "\n",
        "  # The local filename to save the downloaded file\n",
        "  filename = pdf_path\n",
        "\n",
        "  # Send a GET request to the URL\n",
        "  response = requests.get(url)\n",
        "\n",
        "  # Check if the request was successful\n",
        "  if response.status_code == 200:\n",
        "      # Open a file in binary write mode and save the content to it\n",
        "      with open(filename, \"wb\") as file:\n",
        "          file.write(response.content)\n",
        "      print(f\"The file has been downloaded and saved as {filename}\")\n",
        "  else:\n",
        "      print(f\"Failed to download the file. Status code: {response.status_code}\")\n",
        "else:\n",
        "  print(f\"File {pdf_path} exists.\")"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "f9BbUT-Gdt9l",
        "outputId": "7fa29e3c-54b9-4f65-bacd-458dd221e5ad"
      },
      "execution_count": 2,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "File 2205.07690v1.pdf exists.\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# Requires !pip install PyMuPDF, see: https://github.com/pymupdf/pymupdf\n",
        "import fitz # (pymupdf, found this is better than pypdf for our use case, note: licence is AGPL-3.0, keep that in mind if you want to use any code commercially)\n",
        "from tqdm.auto import tqdm # for progress bars, requires !pip install tqdm\n",
        "\n",
        "def text_formatter(text: str) -> str:\n",
        "    \"\"\"Performs minor formatting on text.\"\"\"\n",
        "    cleaned_text = text.replace(\"\\n\", \" \").strip() # note: this might be different for each doc (best to experiment)\n",
        "\n",
        "    # Other potential text formatting functions can go here\n",
        "    return cleaned_text\n",
        "\n",
        "# Open PDF and get lines/pages\n",
        "# Note: this only focuses on text, rather than images/figures etc\n",
        "def open_and_read_pdf(pdf_path: str) -> list[dict]:\n",
        "    \"\"\"\n",
        "    Opens a PDF file, reads its text content page by page, and collects statistics.\n",
        "\n",
        "    Parameters:\n",
        "        pdf_path (str): The file path to the PDF document to be opened and read.\n",
        "\n",
        "    Returns:\n",
        "        list[dict]: A list of dictionaries, each containing the page number\n",
        "        (adjusted), character count, word count, sentence count, token count, and the extracted text\n",
        "        for each page.\n",
        "    \"\"\"\n",
        "    doc = fitz.open(pdf_path)  # open a document\n",
        "    pages_and_texts = []\n",
        "    for page_number, page in tqdm(enumerate(doc)):  # iterate the document pages\n",
        "        text = page.get_text()  # get plain text encoded as UTF-8\n",
        "        text = text_formatter(text)\n",
        "        pages_and_texts.append({\"page_number\": page_number - 41,  # adjust page numbers since our PDF starts on page 42\n",
        "                                \"page_char_count\": len(text),\n",
        "                                \"page_word_count\": len(text.split(\" \")),\n",
        "                                \"page_sentence_count_raw\": len(text.split(\". \")),\n",
        "                                \"page_token_count\": len(text) / 4,  # 1 token = ~4 chars, see: https://help.openai.com/en/articles/4936856-what-are-tokens-and-how-to-count-them\n",
        "                                \"text\": text})\n",
        "    return pages_and_texts\n",
        "\n",
        "pages_and_texts = open_and_read_pdf(pdf_path=pdf_path)\n",
        "pages_and_texts[:2]"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 1000,
          "referenced_widgets": [
            "01a5294ce7d04a2f93d9169e26f4a508",
            "f1bf2f7393804ab98858767a04530e4d",
            "34628447104e4a44a139ddb705b02150",
            "cfd8aa90600241bd97f3161bce9763a6",
            "5c9db98af10f4b3f8f746dab51b6234d",
            "af0a9644c4174d809af0ef4d3dfa7a0c",
            "001856a52d2b4e0e9fd9cb674a329dd5",
            "2c9f5b3371134d0eb9d357f14f576c93",
            "9244e5e7a2ce4a3ebceaf4b21bf60c58",
            "b4452006845d4231b4d56a753b44b0f1",
            "f885417eb70e4e7b9bb92452de2d8a5a"
          ]
        },
        "id": "VZkL9yqVe32N",
        "outputId": "50d7ed7f-2f8b-404b-d12f-4701f3a508e2"
      },
      "execution_count": 3,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "0it [00:00, ?it/s]"
            ],
            "application/vnd.jupyter.widget-view+json": {
              "version_major": 2,
              "version_minor": 0,
              "model_id": "01a5294ce7d04a2f93d9169e26f4a508"
            }
          },
          "metadata": {}
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "[{'page_number': -41,\n",
              "  'page_char_count': 2365,\n",
              "  'page_word_count': 330,\n",
              "  'page_sentence_count_raw': 7,\n",
              "  'page_token_count': 591.25,\n",
              "  'text': 'REAL-TIME SEMANTIC SEGMENTATION ON FPGAS FOR AUTONOMOUS VEHICLES WITH HLS4ML Nicolò Ghielmetti∗, Vladimir Loncar†, Maurizio Pierini, Marcel Roed‡, Sioni Summers European Organization for Nuclear Research (CERN) CH-1211 Geneva 23, Switzerland Thea Aarrestad Institute for Particle Physics and Astrophysics, ETH Zürich 8093 Zürich, Switzerland Christoffer Petersson§ Zenseact Gothenburg, 41756, Sweden Hampus Linander University of Gothenburg Gothenburg, 40530, Sweden Jennifer Ngadiuba Fermi National Accelerator Laboratory Batavia, IL 60510, USA Kelvin Lin¶ University of Washington Seattle, WA 98195, USA Philip Harris Massachusetts Institute of Technology Cambridge, MA 02139, USA May 17, 2022 ABSTRACT In this paper, we investigate how ﬁeld programmable gate arrays can serve as hardware accelerators for real-time semantic segmentation tasks relevant for autonomous driving. Considering compressed versions of the ENet convolutional neural network architecture, we demonstrate a fully-on-chip deployment with a latency of 4.9 ms per image, using less than 30% of the available resources on a Xilinx ZCU102 evaluation board. The latency is reduced to 3 ms per image when increasing the batch size to ten, corresponding to the use case where the autonomous vehicle receives inputs from multiple cameras simultaneously. We show, through aggressive ﬁlter reduction and heterogeneous quantization-aware training, and an optimized implementation of convolutional layers, that the power consumption and resource utilization can be signiﬁcantly reduced while maintaining accuracy on the Cityscapes dataset. Keywords deep learning · FPGA · convolutional neural network 1 Introduction Deep Learning has strongly reshaped computer vision in the last decade, bringing the accuracy of image recognition applications to unprecedented levels. Improved pattern recognition capabilities have had a signiﬁcant impact on the advancement of research in science and technology. Many of the challenges faced by future scientiﬁc experiments, such as the CERN High Luminosity LHC [1] or the Square Kilometer Array observatory [2], and technological ∗Also at Politecnico di Milano, Italy †Also at Institute of Physics Belgrade, Serbia ‡Also at University of Oxford, UK §Also at Chalmers University of Technology, Sweden ¶Currently at Amazon, USA arXiv:2205.07690v1  [cs.CV]  16 May 2022'},\n",
              " {'page_number': -40,\n",
              "  'page_char_count': 5351,\n",
              "  'page_word_count': 813,\n",
              "  'page_sentence_count_raw': 36,\n",
              "  'page_token_count': 1337.75,\n",
              "  'text': 'challenges faced by, for example, the automotive industry, will require the capability of processing large amounts of data in real-time, often through edge computing devices with strict latency and power-consumption constraints. This requirement has generated interest in the development of energy-effective neural networks, resulting in efforts like tinyML [3], which aims to reduce power consumption as much as possible without negatively affecting the model accuracy. Advances in Deep Learning for computer vision have had a crucial impact on the development of autonomous vehicles, enabling the vehicles to perceive their environment at ever-increasing levels of accuracy and detail. Deep Neural Networks are used for ﬁnding patterns and extracting relevant information from camera images, such as the precise location of the surrounding vehicles and pedestrians. In order for an autonomous vehicle to drive safely and efﬁciently, it must be able to react fast and make quick decisions. This imposes strict latency requirements on the neural networks that are deployed to run inference on resource-limited embedded hardware in the vehicle. In addition to algorithmic development, computer vision for autonomous vehicles has beneﬁted from technological advances in parallel computing architecture [4]. The possibility of performing network training and inference on graphics processing units (GPUs) has made large and complex networks computationally affordable and testable on real-life problems. Due to their high efﬁciency, GPUs have become a common hardware choice in the automotive industry for on-vehicle Deep Learning inference. In this paper, we investigate the possibility of exploiting ﬁeld-programmable gate arrays (FPGAs) as a low-power, inference-optimized, highly parallelisable alternative to GPUs. By applying aggressive ﬁlter-reduction and quantization of the model bit precision at training time, and by introducing a highly optimized ﬁrmware implementation of convolutional layers, we achieve the compression required to ﬁt semantic segmentation models on FPGAs. We do so by exploiting and improving the hls4ml library, which provides an automatic conversion of a given Deep Neural Network into C++ code, which is given as input to a high level synthesis (HLS) library. The HLS library then translates this into FPGA ﬁrmware, to be deployed on hardware. Originally developed for scientiﬁc applications in particle physics that require sub-microsecond latency [5–11], hls4ml has been successfully applied outside the domain of scientiﬁc research [12, 13], speciﬁcally in the context of tinyML applications [14]. Applying model compression at training time is crucial in order to minimize resource-consumption and maximize the model accuracy. To do so, we rely on quantization-aware training through the QKeras [15] library, which has been interfaced to hls4ml in order to guarantee an end-to-end optimal training-to-inference workﬂow [12]. As a baseline, we start from the ENet [16] architecture, designed speciﬁcally to perform pixel-wise semantic segmenta- tion for tasks requiring low latency operations. We modify the architecture, removing resource-consuming asymmetric convolutions, and dilated or strided convolutions. In addition, we apply ﬁlter ablation and quantization at training time. Finally, we optimize the implementation of convolutional layers in hls4ml in order to signiﬁcantly reduce the resource consumption. With these steps, we obtain a good balance between resource utilization and accuracy, enabling us to deploy the whole network on a Xilinx ZCU102 evaluation board [17]. This paper is organized as follows: The baseline dataset and model are described in sections 2 and 3, respectively. The model compression and the speciﬁc optimization necessary to port the compressed model to the FPGA are described in section 4 and 5. Conclusions are given in section 6. 2 Dataset Our experiments are performed using the Cityscapes dataset [18], which involves 5,000 trafﬁc scene images collected in 50 different cities with varying road types and seasons. These images have ﬁne-grained semantic segmentation annotations with pixel-level classiﬁcation labels. We have limited ourselves to the four semantic classes Road, Car, Person and Background. According to the standard Cityscapes split, 2975 images are used for training, 500 for validation and 1525 for testing. We crop and resize the original images to have an input resolution of 240×152 pixels. As a pre-processing step, we normalize all pixel values (integer values in the range [0, 255]) to be in the [0, 1] range by dividing each one by 256. In this way all inputs are smaller than one and can be represented by a ﬁxed-point datatype using only 8 bits (log2(256)) (see section 4). An example image from the dataset is shown in Fig. 1, together with a visualization of its semantic segmentation mask. For evaluation metrics we use two typical ﬁgures of merit for semantic segmentation: • The model accuracy (Acc), deﬁned as Acc = T P +T N T P +T N+F P +F N , where TP, TN, FP, and FN are the fraction of true positives, true negatives, false positives, and false negatives, respectively. • The mean of the class-wise Intersection over Union (mIoU), i.e., the average across classes of the Intersection- Over-Union (deﬁned as IOU = T P T P +F P +F N ). 2'}]"
            ]
          },
          "metadata": {},
          "execution_count": 3
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "import random\n",
        "\n",
        "random.sample(pages_and_texts, k=3)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "bkpFgdcrfGCl",
        "outputId": "984662cd-a437-4752-96b4-b36d3eaad242"
      },
      "execution_count": 4,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "[{'page_number': -35,\n",
              "  'page_char_count': 3572,\n",
              "  'page_word_count': 621,\n",
              "  'page_sentence_count_raw': 28,\n",
              "  'page_token_count': 893.0,\n",
              "  'text': 'Input image b a c a a b c Shift registers Sliding input window c b Figure 5: Schematic representation of the new hls4ml implementation of Convolutional layers, as described in the text. The hyperparameter scan is done sequentially over the blocks, i.e. the Bayesian search over quantization and ﬁlter count of the initial layer is performed ﬁrst and is then frozen for the hyperparameter scan of the ﬁrst bottleneck and so on. The rest of the model is kept in ﬂoating point until everything in the end is quantized. Figure 4 shows the outcome of the heterogeneous QAT, in terms of validation accuracy and total number of bits for the six blocks in the network. The optimal conﬁguration search is performed taking as a baseline the Enet4 model, scanning the kernel bits in {4, 8} and ﬁxing the number of kernels to four times a by-layer multiplicative chosen in {0.5, 0.75, 1.0, 1.25, 1.5, 1.75, 2.0}. The optimal conﬁguration (EnetHQ) is obtained for fi = 8, f1 = 2, f2 = 4, f3 = 8, f4 = 4, and f5 = 3, resulting in 4.7 · 103 parameters, a mIoU=36.8% and an accuracy of 81.1%. Out of all the quantized models, both homogeneous and heterogeneous, this is the one which performs the best. 5 FPGA implementation, deployment and results 5.1 Resource-efﬁcient convolution algorithm The hls4ml library has an implementation of convolutional layers that is aimed at low-latency designs [13]. However, this implementation comes at the expense of high resource utilization. This is due to the number of times pixels of the input image are replicated to maintain the state of a sliding input window. For convolutional layers operating on wider images, like in our case, this overhead can be prohibitively large. In order to reduce the resource consumption of the convolutional layers of the model, we introduce a new algorithm that is more resource efﬁcient. The new implementation, dubbed \"line buffer\", uses shift registers to keep track of previously seen pixels. The primary advantage of the line buffer implementation over the previous one is the reduction of the size of the buffer needed to store the replicated pixels. For an image of size H ×W, with a convolution kernel of size K ×L, the line buffer allocates K −1 buffers (chain of shift registers) of depth W for the rows of the image, while the previous implementation allocates K2 buffers of depth K × (W −K + 1) for the elements in the sliding input window. The algorithm is illustrated on Fig. 5. Initially, each new pixel read from the input image stream is pushed into the shift register chain. If the shift register is full, the ﬁrst element will be popped and it will be pushed into the next shift register in chain. The process is repeated for all K −1 shift registers in the chain. The popped pixels are stacked with the input pixel into a column vector and are pushed as the rightmost column of the input window. The pixels popped from the leftmost column of the input window are not used further. In our implementation, the propagation of new pixels through the shift register chain and the insertion into the sliding input window are completed in a single clock cycle, making the implementation as efﬁcient as the existing hls4ml implementation. To compute the output from the populated sliding input window, we rely on the existing routines of hls4ml. We rely on a set of counters to keep track of input window state to know when to produce an output. The algorithm for maintaining the chain of shift registers and populating the sliding input window can be adapted for use in the pooling layers as well. 7'},\n",
              " {'page_number': -36,\n",
              "  'page_char_count': 3057,\n",
              "  'page_word_count': 562,\n",
              "  'page_sentence_count_raw': 18,\n",
              "  'page_token_count': 764.25,\n",
              "  'text': 'Model name Quantization fi f1 f2 f3 f4 f5 mIoU (%) Accuracy (%) Enet16 - 32 16 16 16 16 16 54.3 87.9 Enet8 - 32 8 8 8 8 8 49.4 85.6 Enet4 - 32 4 4 4 4 4 36.6 81.5 Enet16Q8 8 32 16 16 16 16 16 35.0 79.1 Enet8Q8 8 32 8 8 8 8 8 33.4 77.1 Enet4Q8 8 32 4 4 4 4 4 13.6 53.8 Enet16Q4 4 32 16 16 16 16 16 34.1 77.9 Enet8Q4 4 32 8 8 8 8 8 33.9 77.6 Enet4Q4 4 32 4 4 4 4 4 13.5 53.6 Enet16Q2 2 32 16 16 16 16 16 27.4 68.6 Enet8Q2 2 32 8 8 8 8 8 28.7 71.1 Enet4Q2 2 32 4 4 4 4 4 13.4 53.5 EnetHQ Heterogeneous 8 2 4 8 4 3 36.8 81.1 Table 3: Homogeneously and heterogeneously quantized models with indicated bitwidth and ﬁlter architecture together with their validation mean IOU trained with quantization aware training using QKeras. The corresponding values before quantization (from Table 2) are also reported in the three ﬁrst rows. 5.5 6.0 6.5 7.0 total bits 1e6 0.6 0.7 0.8 accuracy Initial block 4 6 8 2 filters 4 filters 6 filters 8 filters 4.8 5.0 5.2 5.4 5.6 5.8 6.0 total bits 1e6 0.76 0.78 0.80 accuracy Block 1 4.7 4.8 4.9 5.0 5.1 total bits 1e6 0.77 0.78 0.79 0.80 accuracy Block 2 4.5 4.6 4.7 4.8 4.9 total bits 1e6 0.77 0.78 0.79 0.80 accuracy Block 3 4.6 4.7 4.8 4.9 5.0 5.1 total bits 1e6 0.78 0.79 0.80 accuracy Block 4 4.5 5.0 5.5 6.0 total bits 1e6 0.6 0.7 0.8 accuracy Block 5 Figure 4: Validation accuracy versus number of bits for the models tested during heterogeneous QAT with AUTO- QKERAS, for the six blocks in the network. The circle size represents the number of ﬁlters, while the color refers to the median bits for the convolutions in the corresponding block. Details on the grid of options considered in the optimization are given in the text. The quantization of the model does not have to be homogeneous across layers. In fact, it has been demonstrated that a heterogeneous quantization is the best way to maintain high accuracy at low resource-cost [22]. We therefore deﬁne one ﬁnal model with an optimized combination of quantizers. 4.3 Heterogeneous quantization aware training Heterogeneous QAT consists in applying different quantization to different network components. For deep networks, one typically deals with the large number of possible conﬁgurations by using an optimization library. In our case, we use AUTOQKERAS [12]. In AUTOQKERAS, a hyperparameter search over individual layer quantization conditions and ﬁlter counts is performed. Since the model contains skip connections, the scan over number of ﬁlters needs to be handled with care. In particular, we use the block features of AUTOQKERAS to ensure that the ﬁlter count matches throughout a bottleneck, so that the tensor addition of the skip connection will have valid dimensions. The search for best hyperparameters, including the choice of indivdual quantizers for kernels and activations, is carried out using a Bayesian strategy where the balance between accuracy and resource usage is controlled by targeting a metric derived from them both[12]. In our search we permit e.g. a 4% decrease in accuracy if the resource usage also is halved at the same time. 6'},\n",
              " {'page_number': -32,\n",
              "  'page_char_count': 359,\n",
              "  'page_word_count': 59,\n",
              "  'page_sentence_count_raw': 14,\n",
              "  'page_token_count': 89.75,\n",
              "  'text': 'M. P., M. R., S. S., and V. L. are supported by the European Research Council (ERC) under the European Union’s Horizon 2020 research and innovation program (Grant Agreement No. 772369). M. P. and N. G. are supported by the European Research Council (ERC) under the European Union’s Horizon 2020 research and innovation program (Grant Agreement No. 966696). 10'}]"
            ]
          },
          "metadata": {},
          "execution_count": 4
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "import pandas as pd\n",
        "\n",
        "df = pd.DataFrame(pages_and_texts)\n",
        "df.head()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 310
        },
        "id": "cEk5Ge0sfPId",
        "outputId": "5e1f18e1-a1ca-40ee-a856-174f2a14ae00"
      },
      "execution_count": 5,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "   page_number  page_char_count  page_word_count  page_sentence_count_raw  \\\n",
              "0          -41             2365              330                        7   \n",
              "1          -40             5351              813                       36   \n",
              "2          -39             3091              508                       24   \n",
              "3          -38             2574              437                       19   \n",
              "4          -37             2274              364                       13   \n",
              "\n",
              "   page_token_count                                               text  \n",
              "0            591.25  REAL-TIME SEMANTIC SEGMENTATION ON FPGAS FOR A...  \n",
              "1           1337.75  challenges faced by, for example, the automoti...  \n",
              "2            772.75  Figure 1: An downsampled image from the Citysc...  \n",
              "3            643.50  Maxpool(2) Pad(2) Concat  Skip   connection Co...  \n",
              "4            568.50  Skip branch Main branch Maxpool(2,2) Pad(1) Co...  "
            ],
            "text/html": [
              "\n",
              "  <div id=\"df-fc8f1106-25c2-401b-aa78-a2e13228197b\" class=\"colab-df-container\">\n",
              "    <div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>page_number</th>\n",
              "      <th>page_char_count</th>\n",
              "      <th>page_word_count</th>\n",
              "      <th>page_sentence_count_raw</th>\n",
              "      <th>page_token_count</th>\n",
              "      <th>text</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>-41</td>\n",
              "      <td>2365</td>\n",
              "      <td>330</td>\n",
              "      <td>7</td>\n",
              "      <td>591.25</td>\n",
              "      <td>REAL-TIME SEMANTIC SEGMENTATION ON FPGAS FOR A...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>-40</td>\n",
              "      <td>5351</td>\n",
              "      <td>813</td>\n",
              "      <td>36</td>\n",
              "      <td>1337.75</td>\n",
              "      <td>challenges faced by, for example, the automoti...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>-39</td>\n",
              "      <td>3091</td>\n",
              "      <td>508</td>\n",
              "      <td>24</td>\n",
              "      <td>772.75</td>\n",
              "      <td>Figure 1: An downsampled image from the Citysc...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>-38</td>\n",
              "      <td>2574</td>\n",
              "      <td>437</td>\n",
              "      <td>19</td>\n",
              "      <td>643.50</td>\n",
              "      <td>Maxpool(2) Pad(2) Concat  Skip   connection Co...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>-37</td>\n",
              "      <td>2274</td>\n",
              "      <td>364</td>\n",
              "      <td>13</td>\n",
              "      <td>568.50</td>\n",
              "      <td>Skip branch Main branch Maxpool(2,2) Pad(1) Co...</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>\n",
              "    <div class=\"colab-df-buttons\">\n",
              "\n",
              "  <div class=\"colab-df-container\">\n",
              "    <button class=\"colab-df-convert\" onclick=\"convertToInteractive('df-fc8f1106-25c2-401b-aa78-a2e13228197b')\"\n",
              "            title=\"Convert this dataframe to an interactive table.\"\n",
              "            style=\"display:none;\">\n",
              "\n",
              "  <svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\" viewBox=\"0 -960 960 960\">\n",
              "    <path d=\"M120-120v-720h720v720H120Zm60-500h600v-160H180v160Zm220 220h160v-160H400v160Zm0 220h160v-160H400v160ZM180-400h160v-160H180v160Zm440 0h160v-160H620v160ZM180-180h160v-160H180v160Zm440 0h160v-160H620v160Z\"/>\n",
              "  </svg>\n",
              "    </button>\n",
              "\n",
              "  <style>\n",
              "    .colab-df-container {\n",
              "      display:flex;\n",
              "      gap: 12px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert {\n",
              "      background-color: #E8F0FE;\n",
              "      border: none;\n",
              "      border-radius: 50%;\n",
              "      cursor: pointer;\n",
              "      display: none;\n",
              "      fill: #1967D2;\n",
              "      height: 32px;\n",
              "      padding: 0 0 0 0;\n",
              "      width: 32px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert:hover {\n",
              "      background-color: #E2EBFA;\n",
              "      box-shadow: 0px 1px 2px rgba(60, 64, 67, 0.3), 0px 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
              "      fill: #174EA6;\n",
              "    }\n",
              "\n",
              "    .colab-df-buttons div {\n",
              "      margin-bottom: 4px;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert {\n",
              "      background-color: #3B4455;\n",
              "      fill: #D2E3FC;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert:hover {\n",
              "      background-color: #434B5C;\n",
              "      box-shadow: 0px 1px 3px 1px rgba(0, 0, 0, 0.15);\n",
              "      filter: drop-shadow(0px 1px 2px rgba(0, 0, 0, 0.3));\n",
              "      fill: #FFFFFF;\n",
              "    }\n",
              "  </style>\n",
              "\n",
              "    <script>\n",
              "      const buttonEl =\n",
              "        document.querySelector('#df-fc8f1106-25c2-401b-aa78-a2e13228197b button.colab-df-convert');\n",
              "      buttonEl.style.display =\n",
              "        google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
              "\n",
              "      async function convertToInteractive(key) {\n",
              "        const element = document.querySelector('#df-fc8f1106-25c2-401b-aa78-a2e13228197b');\n",
              "        const dataTable =\n",
              "          await google.colab.kernel.invokeFunction('convertToInteractive',\n",
              "                                                    [key], {});\n",
              "        if (!dataTable) return;\n",
              "\n",
              "        const docLinkHtml = 'Like what you see? Visit the ' +\n",
              "          '<a target=\"_blank\" href=https://colab.research.google.com/notebooks/data_table.ipynb>data table notebook</a>'\n",
              "          + ' to learn more about interactive tables.';\n",
              "        element.innerHTML = '';\n",
              "        dataTable['output_type'] = 'display_data';\n",
              "        await google.colab.output.renderOutput(dataTable, element);\n",
              "        const docLink = document.createElement('div');\n",
              "        docLink.innerHTML = docLinkHtml;\n",
              "        element.appendChild(docLink);\n",
              "      }\n",
              "    </script>\n",
              "  </div>\n",
              "\n",
              "\n",
              "<div id=\"df-d64cc0a5-8955-4f61-beee-2850a94bc082\">\n",
              "  <button class=\"colab-df-quickchart\" onclick=\"quickchart('df-d64cc0a5-8955-4f61-beee-2850a94bc082')\"\n",
              "            title=\"Suggest charts\"\n",
              "            style=\"display:none;\">\n",
              "\n",
              "<svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\"viewBox=\"0 0 24 24\"\n",
              "     width=\"24px\">\n",
              "    <g>\n",
              "        <path d=\"M19 3H5c-1.1 0-2 .9-2 2v14c0 1.1.9 2 2 2h14c1.1 0 2-.9 2-2V5c0-1.1-.9-2-2-2zM9 17H7v-7h2v7zm4 0h-2V7h2v10zm4 0h-2v-4h2v4z\"/>\n",
              "    </g>\n",
              "</svg>\n",
              "  </button>\n",
              "\n",
              "<style>\n",
              "  .colab-df-quickchart {\n",
              "      --bg-color: #E8F0FE;\n",
              "      --fill-color: #1967D2;\n",
              "      --hover-bg-color: #E2EBFA;\n",
              "      --hover-fill-color: #174EA6;\n",
              "      --disabled-fill-color: #AAA;\n",
              "      --disabled-bg-color: #DDD;\n",
              "  }\n",
              "\n",
              "  [theme=dark] .colab-df-quickchart {\n",
              "      --bg-color: #3B4455;\n",
              "      --fill-color: #D2E3FC;\n",
              "      --hover-bg-color: #434B5C;\n",
              "      --hover-fill-color: #FFFFFF;\n",
              "      --disabled-bg-color: #3B4455;\n",
              "      --disabled-fill-color: #666;\n",
              "  }\n",
              "\n",
              "  .colab-df-quickchart {\n",
              "    background-color: var(--bg-color);\n",
              "    border: none;\n",
              "    border-radius: 50%;\n",
              "    cursor: pointer;\n",
              "    display: none;\n",
              "    fill: var(--fill-color);\n",
              "    height: 32px;\n",
              "    padding: 0;\n",
              "    width: 32px;\n",
              "  }\n",
              "\n",
              "  .colab-df-quickchart:hover {\n",
              "    background-color: var(--hover-bg-color);\n",
              "    box-shadow: 0 1px 2px rgba(60, 64, 67, 0.3), 0 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
              "    fill: var(--button-hover-fill-color);\n",
              "  }\n",
              "\n",
              "  .colab-df-quickchart-complete:disabled,\n",
              "  .colab-df-quickchart-complete:disabled:hover {\n",
              "    background-color: var(--disabled-bg-color);\n",
              "    fill: var(--disabled-fill-color);\n",
              "    box-shadow: none;\n",
              "  }\n",
              "\n",
              "  .colab-df-spinner {\n",
              "    border: 2px solid var(--fill-color);\n",
              "    border-color: transparent;\n",
              "    border-bottom-color: var(--fill-color);\n",
              "    animation:\n",
              "      spin 1s steps(1) infinite;\n",
              "  }\n",
              "\n",
              "  @keyframes spin {\n",
              "    0% {\n",
              "      border-color: transparent;\n",
              "      border-bottom-color: var(--fill-color);\n",
              "      border-left-color: var(--fill-color);\n",
              "    }\n",
              "    20% {\n",
              "      border-color: transparent;\n",
              "      border-left-color: var(--fill-color);\n",
              "      border-top-color: var(--fill-color);\n",
              "    }\n",
              "    30% {\n",
              "      border-color: transparent;\n",
              "      border-left-color: var(--fill-color);\n",
              "      border-top-color: var(--fill-color);\n",
              "      border-right-color: var(--fill-color);\n",
              "    }\n",
              "    40% {\n",
              "      border-color: transparent;\n",
              "      border-right-color: var(--fill-color);\n",
              "      border-top-color: var(--fill-color);\n",
              "    }\n",
              "    60% {\n",
              "      border-color: transparent;\n",
              "      border-right-color: var(--fill-color);\n",
              "    }\n",
              "    80% {\n",
              "      border-color: transparent;\n",
              "      border-right-color: var(--fill-color);\n",
              "      border-bottom-color: var(--fill-color);\n",
              "    }\n",
              "    90% {\n",
              "      border-color: transparent;\n",
              "      border-bottom-color: var(--fill-color);\n",
              "    }\n",
              "  }\n",
              "</style>\n",
              "\n",
              "  <script>\n",
              "    async function quickchart(key) {\n",
              "      const quickchartButtonEl =\n",
              "        document.querySelector('#' + key + ' button');\n",
              "      quickchartButtonEl.disabled = true;  // To prevent multiple clicks.\n",
              "      quickchartButtonEl.classList.add('colab-df-spinner');\n",
              "      try {\n",
              "        const charts = await google.colab.kernel.invokeFunction(\n",
              "            'suggestCharts', [key], {});\n",
              "      } catch (error) {\n",
              "        console.error('Error during call to suggestCharts:', error);\n",
              "      }\n",
              "      quickchartButtonEl.classList.remove('colab-df-spinner');\n",
              "      quickchartButtonEl.classList.add('colab-df-quickchart-complete');\n",
              "    }\n",
              "    (() => {\n",
              "      let quickchartButtonEl =\n",
              "        document.querySelector('#df-d64cc0a5-8955-4f61-beee-2850a94bc082 button');\n",
              "      quickchartButtonEl.style.display =\n",
              "        google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
              "    })();\n",
              "  </script>\n",
              "</div>\n",
              "\n",
              "    </div>\n",
              "  </div>\n"
            ],
            "application/vnd.google.colaboratory.intrinsic+json": {
              "type": "dataframe",
              "variable_name": "df",
              "summary": "{\n  \"name\": \"df\",\n  \"rows\": 11,\n  \"fields\": [\n    {\n      \"column\": \"page_number\",\n      \"properties\": {\n        \"dtype\": \"number\",\n        \"std\": 3,\n        \"min\": -41,\n        \"max\": -31,\n        \"num_unique_values\": 11,\n        \"samples\": [\n          -36,\n          -41,\n          -32\n        ],\n        \"semantic_type\": \"\",\n        \"description\": \"\"\n      }\n    },\n    {\n      \"column\": \"page_char_count\",\n      \"properties\": {\n        \"dtype\": \"number\",\n        \"std\": 1376,\n        \"min\": 359,\n        \"max\": 5351,\n        \"num_unique_values\": 11,\n        \"samples\": [\n          3057,\n          2365,\n          359\n        ],\n        \"semantic_type\": \"\",\n        \"description\": \"\"\n      }\n    },\n    {\n      \"column\": \"page_word_count\",\n      \"properties\": {\n        \"dtype\": \"number\",\n        \"std\": 219,\n        \"min\": 59,\n        \"max\": 814,\n        \"num_unique_values\": 11,\n        \"samples\": [\n          562,\n          330,\n          59\n        ],\n        \"semantic_type\": \"\",\n        \"description\": \"\"\n      }\n    },\n    {\n      \"column\": \"page_sentence_count_raw\",\n      \"properties\": {\n        \"dtype\": \"number\",\n        \"std\": 25,\n        \"min\": 7,\n        \"max\": 101,\n        \"num_unique_values\": 11,\n        \"samples\": [\n          18,\n          7,\n          14\n        ],\n        \"semantic_type\": \"\",\n        \"description\": \"\"\n      }\n    },\n    {\n      \"column\": \"page_token_count\",\n      \"properties\": {\n        \"dtype\": \"number\",\n        \"std\": 344.0714678667791,\n        \"min\": 89.75,\n        \"max\": 1337.75,\n        \"num_unique_values\": 11,\n        \"samples\": [\n          764.25,\n          591.25,\n          89.75\n        ],\n        \"semantic_type\": \"\",\n        \"description\": \"\"\n      }\n    },\n    {\n      \"column\": \"text\",\n      \"properties\": {\n        \"dtype\": \"string\",\n        \"num_unique_values\": 11,\n        \"samples\": [\n          \"Model name Quantization fi f1 f2 f3 f4 f5 mIoU (%) Accuracy (%) Enet16 - 32 16 16 16 16 16 54.3 87.9 Enet8 - 32 8 8 8 8 8 49.4 85.6 Enet4 - 32 4 4 4 4 4 36.6 81.5 Enet16Q8 8 32 16 16 16 16 16 35.0 79.1 Enet8Q8 8 32 8 8 8 8 8 33.4 77.1 Enet4Q8 8 32 4 4 4 4 4 13.6 53.8 Enet16Q4 4 32 16 16 16 16 16 34.1 77.9 Enet8Q4 4 32 8 8 8 8 8 33.9 77.6 Enet4Q4 4 32 4 4 4 4 4 13.5 53.6 Enet16Q2 2 32 16 16 16 16 16 27.4 68.6 Enet8Q2 2 32 8 8 8 8 8 28.7 71.1 Enet4Q2 2 32 4 4 4 4 4 13.4 53.5 EnetHQ Heterogeneous 8 2 4 8 4 3 36.8 81.1 Table 3: Homogeneously and heterogeneously quantized models with indicated bitwidth and \\ufb01lter architecture together with their validation mean IOU trained with quantization aware training using QKeras. The corresponding values before quantization (from Table 2) are also reported in the three \\ufb01rst rows. 5.5 6.0 6.5 7.0 total bits 1e6 0.6 0.7 0.8 accuracy Initial block 4 6 8 2 filters 4 filters 6 filters 8 filters 4.8 5.0 5.2 5.4 5.6 5.8 6.0 total bits 1e6 0.76 0.78 0.80 accuracy Block 1 4.7 4.8 4.9 5.0 5.1 total bits 1e6 0.77 0.78 0.79 0.80 accuracy Block 2 4.5 4.6 4.7 4.8 4.9 total bits 1e6 0.77 0.78 0.79 0.80 accuracy Block 3 4.6 4.7 4.8 4.9 5.0 5.1 total bits 1e6 0.78 0.79 0.80 accuracy Block 4 4.5 5.0 5.5 6.0 total bits 1e6 0.6 0.7 0.8 accuracy Block 5 Figure 4: Validation accuracy versus number of bits for the models tested during heterogeneous QAT with AUTO- QKERAS, for the six blocks in the network. The circle size represents the number of \\ufb01lters, while the color refers to the median bits for the convolutions in the corresponding block. Details on the grid of options considered in the optimization are given in the text. The quantization of the model does not have to be homogeneous across layers. In fact, it has been demonstrated that a heterogeneous quantization is the best way to maintain high accuracy at low resource-cost [22]. We therefore de\\ufb01ne one \\ufb01nal model with an optimized combination of quantizers. 4.3 Heterogeneous quantization aware training Heterogeneous QAT consists in applying different quantization to different network components. For deep networks, one typically deals with the large number of possible con\\ufb01gurations by using an optimization library. In our case, we use AUTOQKERAS [12]. In AUTOQKERAS, a hyperparameter search over individual layer quantization conditions and \\ufb01lter counts is performed. Since the model contains skip connections, the scan over number of \\ufb01lters needs to be handled with care. In particular, we use the block features of AUTOQKERAS to ensure that the \\ufb01lter count matches throughout a bottleneck, so that the tensor addition of the skip connection will have valid dimensions. The search for best hyperparameters, including the choice of indivdual quantizers for kernels and activations, is carried out using a Bayesian strategy where the balance between accuracy and resource usage is controlled by targeting a metric derived from them both[12]. In our search we permit e.g. a 4% decrease in accuracy if the resource usage also is halved at the same time. 6\",\n          \"REAL-TIME SEMANTIC SEGMENTATION ON FPGAS FOR AUTONOMOUS VEHICLES WITH HLS4ML Nicol\\u00f2 Ghielmetti\\u2217, Vladimir Loncar\\u2020, Maurizio Pierini, Marcel Roed\\u2021, Sioni Summers European Organization for Nuclear Research (CERN) CH-1211 Geneva 23, Switzerland Thea Aarrestad Institute for Particle Physics and Astrophysics, ETH Z\\u00fcrich 8093 Z\\u00fcrich, Switzerland Christoffer Petersson\\u00a7 Zenseact Gothenburg, 41756, Sweden Hampus Linander University of Gothenburg Gothenburg, 40530, Sweden Jennifer Ngadiuba Fermi National Accelerator Laboratory Batavia, IL 60510, USA Kelvin Lin\\u00b6 University of Washington Seattle, WA 98195, USA Philip Harris Massachusetts Institute of Technology Cambridge, MA 02139, USA May 17, 2022 ABSTRACT In this paper, we investigate how \\ufb01eld programmable gate arrays can serve as hardware accelerators for real-time semantic segmentation tasks relevant for autonomous driving. Considering compressed versions of the ENet convolutional neural network architecture, we demonstrate a fully-on-chip deployment with a latency of 4.9 ms per image, using less than 30% of the available resources on a Xilinx ZCU102 evaluation board. The latency is reduced to 3 ms per image when increasing the batch size to ten, corresponding to the use case where the autonomous vehicle receives inputs from multiple cameras simultaneously. We show, through aggressive \\ufb01lter reduction and heterogeneous quantization-aware training, and an optimized implementation of convolutional layers, that the power consumption and resource utilization can be signi\\ufb01cantly reduced while maintaining accuracy on the Cityscapes dataset. Keywords deep learning \\u00b7 FPGA \\u00b7 convolutional neural network 1 Introduction Deep Learning has strongly reshaped computer vision in the last decade, bringing the accuracy of image recognition applications to unprecedented levels. Improved pattern recognition capabilities have had a signi\\ufb01cant impact on the advancement of research in science and technology. Many of the challenges faced by future scienti\\ufb01c experiments, such as the CERN High Luminosity LHC [1] or the Square Kilometer Array observatory [2], and technological \\u2217Also at Politecnico di Milano, Italy \\u2020Also at Institute of Physics Belgrade, Serbia \\u2021Also at University of Oxford, UK \\u00a7Also at Chalmers University of Technology, Sweden \\u00b6Currently at Amazon, USA arXiv:2205.07690v1  [cs.CV]  16 May 2022\",\n          \"M. P., M. R., S. S., and V. L. are supported by the European Research Council (ERC) under the European Union\\u2019s Horizon 2020 research and innovation program (Grant Agreement No. 772369). M. P. and N. G. are supported by the European Research Council (ERC) under the European Union\\u2019s Horizon 2020 research and innovation program (Grant Agreement No. 966696). 10\"\n        ],\n        \"semantic_type\": \"\",\n        \"description\": \"\"\n      }\n    }\n  ]\n}"
            }
          },
          "metadata": {},
          "execution_count": 5
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# Get stats\n",
        "df.describe().round(2)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 300
        },
        "id": "pMg97Zp_fWmi",
        "outputId": "1792d045-3c81-4819-9e66-920096e9e598"
      },
      "execution_count": 6,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "       page_number  page_char_count  page_word_count  page_sentence_count_raw  \\\n",
              "count        11.00            11.00            11.00                    11.00   \n",
              "mean        -36.00          3227.00           515.82                    30.55   \n",
              "std           3.32          1376.29           219.43                    25.75   \n",
              "min         -41.00           359.00            59.00                     7.00   \n",
              "25%         -38.50          2469.50           400.50                    16.00   \n",
              "50%         -36.00          3091.00           514.00                    24.00   \n",
              "75%         -33.50          3967.50           636.50                    34.50   \n",
              "max         -31.00          5351.00           814.00                   101.00   \n",
              "\n",
              "       page_token_count  \n",
              "count             11.00  \n",
              "mean             806.75  \n",
              "std              344.07  \n",
              "min               89.75  \n",
              "25%              617.38  \n",
              "50%              772.75  \n",
              "75%              991.88  \n",
              "max             1337.75  "
            ],
            "text/html": [
              "\n",
              "  <div id=\"df-2f991bd6-f091-4929-9feb-c6fabd9d932a\" class=\"colab-df-container\">\n",
              "    <div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>page_number</th>\n",
              "      <th>page_char_count</th>\n",
              "      <th>page_word_count</th>\n",
              "      <th>page_sentence_count_raw</th>\n",
              "      <th>page_token_count</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>count</th>\n",
              "      <td>11.00</td>\n",
              "      <td>11.00</td>\n",
              "      <td>11.00</td>\n",
              "      <td>11.00</td>\n",
              "      <td>11.00</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>mean</th>\n",
              "      <td>-36.00</td>\n",
              "      <td>3227.00</td>\n",
              "      <td>515.82</td>\n",
              "      <td>30.55</td>\n",
              "      <td>806.75</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>std</th>\n",
              "      <td>3.32</td>\n",
              "      <td>1376.29</td>\n",
              "      <td>219.43</td>\n",
              "      <td>25.75</td>\n",
              "      <td>344.07</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>min</th>\n",
              "      <td>-41.00</td>\n",
              "      <td>359.00</td>\n",
              "      <td>59.00</td>\n",
              "      <td>7.00</td>\n",
              "      <td>89.75</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>25%</th>\n",
              "      <td>-38.50</td>\n",
              "      <td>2469.50</td>\n",
              "      <td>400.50</td>\n",
              "      <td>16.00</td>\n",
              "      <td>617.38</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>50%</th>\n",
              "      <td>-36.00</td>\n",
              "      <td>3091.00</td>\n",
              "      <td>514.00</td>\n",
              "      <td>24.00</td>\n",
              "      <td>772.75</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>75%</th>\n",
              "      <td>-33.50</td>\n",
              "      <td>3967.50</td>\n",
              "      <td>636.50</td>\n",
              "      <td>34.50</td>\n",
              "      <td>991.88</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>max</th>\n",
              "      <td>-31.00</td>\n",
              "      <td>5351.00</td>\n",
              "      <td>814.00</td>\n",
              "      <td>101.00</td>\n",
              "      <td>1337.75</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>\n",
              "    <div class=\"colab-df-buttons\">\n",
              "\n",
              "  <div class=\"colab-df-container\">\n",
              "    <button class=\"colab-df-convert\" onclick=\"convertToInteractive('df-2f991bd6-f091-4929-9feb-c6fabd9d932a')\"\n",
              "            title=\"Convert this dataframe to an interactive table.\"\n",
              "            style=\"display:none;\">\n",
              "\n",
              "  <svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\" viewBox=\"0 -960 960 960\">\n",
              "    <path d=\"M120-120v-720h720v720H120Zm60-500h600v-160H180v160Zm220 220h160v-160H400v160Zm0 220h160v-160H400v160ZM180-400h160v-160H180v160Zm440 0h160v-160H620v160ZM180-180h160v-160H180v160Zm440 0h160v-160H620v160Z\"/>\n",
              "  </svg>\n",
              "    </button>\n",
              "\n",
              "  <style>\n",
              "    .colab-df-container {\n",
              "      display:flex;\n",
              "      gap: 12px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert {\n",
              "      background-color: #E8F0FE;\n",
              "      border: none;\n",
              "      border-radius: 50%;\n",
              "      cursor: pointer;\n",
              "      display: none;\n",
              "      fill: #1967D2;\n",
              "      height: 32px;\n",
              "      padding: 0 0 0 0;\n",
              "      width: 32px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert:hover {\n",
              "      background-color: #E2EBFA;\n",
              "      box-shadow: 0px 1px 2px rgba(60, 64, 67, 0.3), 0px 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
              "      fill: #174EA6;\n",
              "    }\n",
              "\n",
              "    .colab-df-buttons div {\n",
              "      margin-bottom: 4px;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert {\n",
              "      background-color: #3B4455;\n",
              "      fill: #D2E3FC;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert:hover {\n",
              "      background-color: #434B5C;\n",
              "      box-shadow: 0px 1px 3px 1px rgba(0, 0, 0, 0.15);\n",
              "      filter: drop-shadow(0px 1px 2px rgba(0, 0, 0, 0.3));\n",
              "      fill: #FFFFFF;\n",
              "    }\n",
              "  </style>\n",
              "\n",
              "    <script>\n",
              "      const buttonEl =\n",
              "        document.querySelector('#df-2f991bd6-f091-4929-9feb-c6fabd9d932a button.colab-df-convert');\n",
              "      buttonEl.style.display =\n",
              "        google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
              "\n",
              "      async function convertToInteractive(key) {\n",
              "        const element = document.querySelector('#df-2f991bd6-f091-4929-9feb-c6fabd9d932a');\n",
              "        const dataTable =\n",
              "          await google.colab.kernel.invokeFunction('convertToInteractive',\n",
              "                                                    [key], {});\n",
              "        if (!dataTable) return;\n",
              "\n",
              "        const docLinkHtml = 'Like what you see? Visit the ' +\n",
              "          '<a target=\"_blank\" href=https://colab.research.google.com/notebooks/data_table.ipynb>data table notebook</a>'\n",
              "          + ' to learn more about interactive tables.';\n",
              "        element.innerHTML = '';\n",
              "        dataTable['output_type'] = 'display_data';\n",
              "        await google.colab.output.renderOutput(dataTable, element);\n",
              "        const docLink = document.createElement('div');\n",
              "        docLink.innerHTML = docLinkHtml;\n",
              "        element.appendChild(docLink);\n",
              "      }\n",
              "    </script>\n",
              "  </div>\n",
              "\n",
              "\n",
              "<div id=\"df-fd2a5580-f7fd-4ab2-a2ac-8ad4b9546f89\">\n",
              "  <button class=\"colab-df-quickchart\" onclick=\"quickchart('df-fd2a5580-f7fd-4ab2-a2ac-8ad4b9546f89')\"\n",
              "            title=\"Suggest charts\"\n",
              "            style=\"display:none;\">\n",
              "\n",
              "<svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\"viewBox=\"0 0 24 24\"\n",
              "     width=\"24px\">\n",
              "    <g>\n",
              "        <path d=\"M19 3H5c-1.1 0-2 .9-2 2v14c0 1.1.9 2 2 2h14c1.1 0 2-.9 2-2V5c0-1.1-.9-2-2-2zM9 17H7v-7h2v7zm4 0h-2V7h2v10zm4 0h-2v-4h2v4z\"/>\n",
              "    </g>\n",
              "</svg>\n",
              "  </button>\n",
              "\n",
              "<style>\n",
              "  .colab-df-quickchart {\n",
              "      --bg-color: #E8F0FE;\n",
              "      --fill-color: #1967D2;\n",
              "      --hover-bg-color: #E2EBFA;\n",
              "      --hover-fill-color: #174EA6;\n",
              "      --disabled-fill-color: #AAA;\n",
              "      --disabled-bg-color: #DDD;\n",
              "  }\n",
              "\n",
              "  [theme=dark] .colab-df-quickchart {\n",
              "      --bg-color: #3B4455;\n",
              "      --fill-color: #D2E3FC;\n",
              "      --hover-bg-color: #434B5C;\n",
              "      --hover-fill-color: #FFFFFF;\n",
              "      --disabled-bg-color: #3B4455;\n",
              "      --disabled-fill-color: #666;\n",
              "  }\n",
              "\n",
              "  .colab-df-quickchart {\n",
              "    background-color: var(--bg-color);\n",
              "    border: none;\n",
              "    border-radius: 50%;\n",
              "    cursor: pointer;\n",
              "    display: none;\n",
              "    fill: var(--fill-color);\n",
              "    height: 32px;\n",
              "    padding: 0;\n",
              "    width: 32px;\n",
              "  }\n",
              "\n",
              "  .colab-df-quickchart:hover {\n",
              "    background-color: var(--hover-bg-color);\n",
              "    box-shadow: 0 1px 2px rgba(60, 64, 67, 0.3), 0 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
              "    fill: var(--button-hover-fill-color);\n",
              "  }\n",
              "\n",
              "  .colab-df-quickchart-complete:disabled,\n",
              "  .colab-df-quickchart-complete:disabled:hover {\n",
              "    background-color: var(--disabled-bg-color);\n",
              "    fill: var(--disabled-fill-color);\n",
              "    box-shadow: none;\n",
              "  }\n",
              "\n",
              "  .colab-df-spinner {\n",
              "    border: 2px solid var(--fill-color);\n",
              "    border-color: transparent;\n",
              "    border-bottom-color: var(--fill-color);\n",
              "    animation:\n",
              "      spin 1s steps(1) infinite;\n",
              "  }\n",
              "\n",
              "  @keyframes spin {\n",
              "    0% {\n",
              "      border-color: transparent;\n",
              "      border-bottom-color: var(--fill-color);\n",
              "      border-left-color: var(--fill-color);\n",
              "    }\n",
              "    20% {\n",
              "      border-color: transparent;\n",
              "      border-left-color: var(--fill-color);\n",
              "      border-top-color: var(--fill-color);\n",
              "    }\n",
              "    30% {\n",
              "      border-color: transparent;\n",
              "      border-left-color: var(--fill-color);\n",
              "      border-top-color: var(--fill-color);\n",
              "      border-right-color: var(--fill-color);\n",
              "    }\n",
              "    40% {\n",
              "      border-color: transparent;\n",
              "      border-right-color: var(--fill-color);\n",
              "      border-top-color: var(--fill-color);\n",
              "    }\n",
              "    60% {\n",
              "      border-color: transparent;\n",
              "      border-right-color: var(--fill-color);\n",
              "    }\n",
              "    80% {\n",
              "      border-color: transparent;\n",
              "      border-right-color: var(--fill-color);\n",
              "      border-bottom-color: var(--fill-color);\n",
              "    }\n",
              "    90% {\n",
              "      border-color: transparent;\n",
              "      border-bottom-color: var(--fill-color);\n",
              "    }\n",
              "  }\n",
              "</style>\n",
              "\n",
              "  <script>\n",
              "    async function quickchart(key) {\n",
              "      const quickchartButtonEl =\n",
              "        document.querySelector('#' + key + ' button');\n",
              "      quickchartButtonEl.disabled = true;  // To prevent multiple clicks.\n",
              "      quickchartButtonEl.classList.add('colab-df-spinner');\n",
              "      try {\n",
              "        const charts = await google.colab.kernel.invokeFunction(\n",
              "            'suggestCharts', [key], {});\n",
              "      } catch (error) {\n",
              "        console.error('Error during call to suggestCharts:', error);\n",
              "      }\n",
              "      quickchartButtonEl.classList.remove('colab-df-spinner');\n",
              "      quickchartButtonEl.classList.add('colab-df-quickchart-complete');\n",
              "    }\n",
              "    (() => {\n",
              "      let quickchartButtonEl =\n",
              "        document.querySelector('#df-fd2a5580-f7fd-4ab2-a2ac-8ad4b9546f89 button');\n",
              "      quickchartButtonEl.style.display =\n",
              "        google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
              "    })();\n",
              "  </script>\n",
              "</div>\n",
              "\n",
              "    </div>\n",
              "  </div>\n"
            ],
            "application/vnd.google.colaboratory.intrinsic+json": {
              "type": "dataframe",
              "summary": "{\n  \"name\": \"df\",\n  \"rows\": 8,\n  \"fields\": [\n    {\n      \"column\": \"page_number\",\n      \"properties\": {\n        \"dtype\": \"number\",\n        \"std\": 20.30541658910604,\n        \"min\": -41.0,\n        \"max\": 11.0,\n        \"num_unique_values\": 7,\n        \"samples\": [\n          11.0,\n          -36.0,\n          -33.5\n        ],\n        \"semantic_type\": \"\",\n        \"description\": \"\"\n      }\n    },\n    {\n      \"column\": \"page_char_count\",\n      \"properties\": {\n        \"dtype\": \"number\",\n        \"std\": 1820.4982200166878,\n        \"min\": 11.0,\n        \"max\": 5351.0,\n        \"num_unique_values\": 8,\n        \"samples\": [\n          3227.0,\n          3091.0,\n          11.0\n        ],\n        \"semantic_type\": \"\",\n        \"description\": \"\"\n      }\n    },\n    {\n      \"column\": \"page_word_count\",\n      \"properties\": {\n        \"dtype\": \"number\",\n        \"std\": 281.2816749115936,\n        \"min\": 11.0,\n        \"max\": 814.0,\n        \"num_unique_values\": 8,\n        \"samples\": [\n          515.82,\n          514.0,\n          11.0\n        ],\n        \"semantic_type\": \"\",\n        \"description\": \"\"\n      }\n    },\n    {\n      \"column\": \"page_sentence_count_raw\",\n      \"properties\": {\n        \"dtype\": \"number\",\n        \"std\": 29.736005111648744,\n        \"min\": 7.0,\n        \"max\": 101.0,\n        \"num_unique_values\": 8,\n        \"samples\": [\n          30.55,\n          24.0,\n          11.0\n        ],\n        \"semantic_type\": \"\",\n        \"description\": \"\"\n      }\n    },\n    {\n      \"column\": \"page_token_count\",\n      \"properties\": {\n        \"dtype\": \"number\",\n        \"std\": 453.532510661348,\n        \"min\": 11.0,\n        \"max\": 1337.75,\n        \"num_unique_values\": 8,\n        \"samples\": [\n          806.75,\n          772.75,\n          11.0\n        ],\n        \"semantic_type\": \"\",\n        \"description\": \"\"\n      }\n    }\n  ]\n}"
            }
          },
          "metadata": {},
          "execution_count": 6
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "from spacy.lang.en import English # see https://spacy.io/usage for install instructions\n",
        "\n",
        "nlp = English()\n",
        "\n",
        "# Add a sentencizer pipeline, see https://spacy.io/api/sentencizer/\n",
        "nlp.add_pipe(\"sentencizer\")\n",
        "\n",
        "# Create a document instance as an example\n",
        "doc = nlp(\"This is a sentence. This another sentence.\")\n",
        "assert len(list(doc.sents)) == 2\n",
        "\n",
        "# Access the sentences of the document\n",
        "list(doc.sents)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "BWkxCi74fgNq",
        "outputId": "a9968815-145f-476b-bfb9-56040cf73505"
      },
      "execution_count": 7,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "[This is a sentence., This another sentence.]"
            ]
          },
          "metadata": {},
          "execution_count": 7
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "for item in tqdm(pages_and_texts):\n",
        "    item[\"sentences\"] = list(nlp(item[\"text\"]).sents)\n",
        "\n",
        "    # Make sure all sentences are strings\n",
        "    item[\"sentences\"] = [str(sentence) for sentence in item[\"sentences\"]]\n",
        "\n",
        "    # Count the sentences\n",
        "    item[\"page_sentence_count_spacy\"] = len(item[\"sentences\"])"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 49,
          "referenced_widgets": [
            "da8bf76154664474b489c8b1c118c732",
            "322abf6001f34aef8fd9aef4f3e4d9ef",
            "d35d104d80f44614a3d5095e419a76ea",
            "95d09829a7784ffb98b7b8012096c719",
            "7b96a8f4b91c4dcf92c2d470b927cda5",
            "180ab15da4c545be9d48a7f358c3dfe2",
            "92b34f04a0044697a9cdd2d617858cd8",
            "9448abe0d444492881e132c60f0ad15b",
            "c9a1c923f8f746d8bfd7714df9dd7ba2",
            "489a7947f1b04f6787bd8c2b1eb19c1c",
            "bd90dd854a22421a9d3e51bf078558aa"
          ]
        },
        "id": "lemqjwklfnBN",
        "outputId": "d1ad0d73-6399-4cd0-d8ee-56cf6262916c"
      },
      "execution_count": 8,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "  0%|          | 0/11 [00:00<?, ?it/s]"
            ],
            "application/vnd.jupyter.widget-view+json": {
              "version_major": 2,
              "version_minor": 0,
              "model_id": "da8bf76154664474b489c8b1c118c732"
            }
          },
          "metadata": {}
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# Inspect an example\n",
        "random.sample(pages_and_texts, k=1)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "_0ecJrUifssn",
        "outputId": "b05c1525-bbc0-435b-f33e-46bf3ecbb362"
      },
      "execution_count": 9,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "[{'page_number': -40,\n",
              "  'page_char_count': 5351,\n",
              "  'page_word_count': 813,\n",
              "  'page_sentence_count_raw': 36,\n",
              "  'page_token_count': 1337.75,\n",
              "  'text': 'challenges faced by, for example, the automotive industry, will require the capability of processing large amounts of data in real-time, often through edge computing devices with strict latency and power-consumption constraints. This requirement has generated interest in the development of energy-effective neural networks, resulting in efforts like tinyML [3], which aims to reduce power consumption as much as possible without negatively affecting the model accuracy. Advances in Deep Learning for computer vision have had a crucial impact on the development of autonomous vehicles, enabling the vehicles to perceive their environment at ever-increasing levels of accuracy and detail. Deep Neural Networks are used for ﬁnding patterns and extracting relevant information from camera images, such as the precise location of the surrounding vehicles and pedestrians. In order for an autonomous vehicle to drive safely and efﬁciently, it must be able to react fast and make quick decisions. This imposes strict latency requirements on the neural networks that are deployed to run inference on resource-limited embedded hardware in the vehicle. In addition to algorithmic development, computer vision for autonomous vehicles has beneﬁted from technological advances in parallel computing architecture [4]. The possibility of performing network training and inference on graphics processing units (GPUs) has made large and complex networks computationally affordable and testable on real-life problems. Due to their high efﬁciency, GPUs have become a common hardware choice in the automotive industry for on-vehicle Deep Learning inference. In this paper, we investigate the possibility of exploiting ﬁeld-programmable gate arrays (FPGAs) as a low-power, inference-optimized, highly parallelisable alternative to GPUs. By applying aggressive ﬁlter-reduction and quantization of the model bit precision at training time, and by introducing a highly optimized ﬁrmware implementation of convolutional layers, we achieve the compression required to ﬁt semantic segmentation models on FPGAs. We do so by exploiting and improving the hls4ml library, which provides an automatic conversion of a given Deep Neural Network into C++ code, which is given as input to a high level synthesis (HLS) library. The HLS library then translates this into FPGA ﬁrmware, to be deployed on hardware. Originally developed for scientiﬁc applications in particle physics that require sub-microsecond latency [5–11], hls4ml has been successfully applied outside the domain of scientiﬁc research [12, 13], speciﬁcally in the context of tinyML applications [14]. Applying model compression at training time is crucial in order to minimize resource-consumption and maximize the model accuracy. To do so, we rely on quantization-aware training through the QKeras [15] library, which has been interfaced to hls4ml in order to guarantee an end-to-end optimal training-to-inference workﬂow [12]. As a baseline, we start from the ENet [16] architecture, designed speciﬁcally to perform pixel-wise semantic segmenta- tion for tasks requiring low latency operations. We modify the architecture, removing resource-consuming asymmetric convolutions, and dilated or strided convolutions. In addition, we apply ﬁlter ablation and quantization at training time. Finally, we optimize the implementation of convolutional layers in hls4ml in order to signiﬁcantly reduce the resource consumption. With these steps, we obtain a good balance between resource utilization and accuracy, enabling us to deploy the whole network on a Xilinx ZCU102 evaluation board [17]. This paper is organized as follows: The baseline dataset and model are described in sections 2 and 3, respectively. The model compression and the speciﬁc optimization necessary to port the compressed model to the FPGA are described in section 4 and 5. Conclusions are given in section 6. 2 Dataset Our experiments are performed using the Cityscapes dataset [18], which involves 5,000 trafﬁc scene images collected in 50 different cities with varying road types and seasons. These images have ﬁne-grained semantic segmentation annotations with pixel-level classiﬁcation labels. We have limited ourselves to the four semantic classes Road, Car, Person and Background. According to the standard Cityscapes split, 2975 images are used for training, 500 for validation and 1525 for testing. We crop and resize the original images to have an input resolution of 240×152 pixels. As a pre-processing step, we normalize all pixel values (integer values in the range [0, 255]) to be in the [0, 1] range by dividing each one by 256. In this way all inputs are smaller than one and can be represented by a ﬁxed-point datatype using only 8 bits (log2(256)) (see section 4). An example image from the dataset is shown in Fig. 1, together with a visualization of its semantic segmentation mask. For evaluation metrics we use two typical ﬁgures of merit for semantic segmentation: • The model accuracy (Acc), deﬁned as Acc = T P +T N T P +T N+F P +F N , where TP, TN, FP, and FN are the fraction of true positives, true negatives, false positives, and false negatives, respectively. • The mean of the class-wise Intersection over Union (mIoU), i.e., the average across classes of the Intersection- Over-Union (deﬁned as IOU = T P T P +F P +F N ). 2',\n",
              "  'sentences': ['challenges faced by, for example, the automotive industry, will require the capability of processing large amounts of data in real-time, often through edge computing devices with strict latency and power-consumption constraints.',\n",
              "   'This requirement has generated interest in the development of energy-effective neural networks, resulting in efforts like tinyML [3], which aims to reduce power consumption as much as possible without negatively affecting the model accuracy.',\n",
              "   'Advances in Deep Learning for computer vision have had a crucial impact on the development of autonomous vehicles, enabling the vehicles to perceive their environment at ever-increasing levels of accuracy and detail.',\n",
              "   'Deep Neural Networks are used for ﬁnding patterns and extracting relevant information from camera images, such as the precise location of the surrounding vehicles and pedestrians.',\n",
              "   'In order for an autonomous vehicle to drive safely and efﬁciently, it must be able to react fast and make quick decisions.',\n",
              "   'This imposes strict latency requirements on the neural networks that are deployed to run inference on resource-limited embedded hardware in the vehicle.',\n",
              "   'In addition to algorithmic development, computer vision for autonomous vehicles has beneﬁted from technological advances in parallel computing architecture [4].',\n",
              "   'The possibility of performing network training and inference on graphics processing units (GPUs) has made large and complex networks computationally affordable and testable on real-life problems.',\n",
              "   'Due to their high efﬁciency, GPUs have become a common hardware choice in the automotive industry for on-vehicle Deep Learning inference.',\n",
              "   'In this paper, we investigate the possibility of exploiting ﬁeld-programmable gate arrays (FPGAs) as a low-power, inference-optimized, highly parallelisable alternative to GPUs.',\n",
              "   'By applying aggressive ﬁlter-reduction and quantization of the model bit precision at training time, and by introducing a highly optimized ﬁrmware implementation of convolutional layers, we achieve the compression required to ﬁt semantic segmentation models on FPGAs.',\n",
              "   'We do so by exploiting and improving the hls4ml library, which provides an automatic conversion of a given Deep Neural Network into C++ code, which is given as input to a high level synthesis (HLS) library.',\n",
              "   'The HLS library then translates this into FPGA ﬁrmware, to be deployed on hardware.',\n",
              "   'Originally developed for scientiﬁc applications in particle physics that require sub-microsecond latency [5–11], hls4ml has been successfully applied outside the domain of scientiﬁc research [12, 13], speciﬁcally in the context of tinyML applications [14].',\n",
              "   'Applying model compression at training time is crucial in order to minimize resource-consumption and maximize the model accuracy.',\n",
              "   'To do so, we rely on quantization-aware training through the QKeras [15] library, which has been interfaced to hls4ml in order to guarantee an end-to-end optimal training-to-inference workﬂow [12].',\n",
              "   'As a baseline, we start from the ENet [16] architecture, designed speciﬁcally to perform pixel-wise semantic segmenta- tion for tasks requiring low latency operations.',\n",
              "   'We modify the architecture, removing resource-consuming asymmetric convolutions, and dilated or strided convolutions.',\n",
              "   'In addition, we apply ﬁlter ablation and quantization at training time.',\n",
              "   'Finally, we optimize the implementation of convolutional layers in hls4ml in order to signiﬁcantly reduce the resource consumption.',\n",
              "   'With these steps, we obtain a good balance between resource utilization and accuracy, enabling us to deploy the whole network on a Xilinx ZCU102 evaluation board [17].',\n",
              "   'This paper is organized as follows: The baseline dataset and model are described in sections 2 and 3, respectively.',\n",
              "   'The model compression and the speciﬁc optimization necessary to port the compressed model to the FPGA are described in section 4 and 5.',\n",
              "   'Conclusions are given in section 6.',\n",
              "   '2 Dataset Our experiments are performed using the Cityscapes dataset [18], which involves 5,000 trafﬁc scene images collected in 50 different cities with varying road types and seasons.',\n",
              "   'These images have ﬁne-grained semantic segmentation annotations with pixel-level classiﬁcation labels.',\n",
              "   'We have limited ourselves to the four semantic classes Road, Car, Person and Background.',\n",
              "   'According to the standard Cityscapes split, 2975 images are used for training, 500 for validation and 1525 for testing.',\n",
              "   'We crop and resize the original images to have an input resolution of 240×152 pixels.',\n",
              "   'As a pre-processing step, we normalize all pixel values (integer values in the range [0, 255]) to be in the [0, 1] range by dividing each one by 256.',\n",
              "   'In this way all inputs are smaller than one and can be represented by a ﬁxed-point datatype using only 8 bits (log2(256)) (see section 4).',\n",
              "   'An example image from the dataset is shown in Fig.',\n",
              "   '1, together with a visualization of its semantic segmentation mask.',\n",
              "   'For evaluation metrics we use two typical ﬁgures of merit for semantic segmentation: • The model accuracy (Acc), deﬁned as Acc = T P +T N T P +T N+F P +F N , where TP, TN, FP, and FN are the fraction of true positives, true negatives, false positives, and false negatives, respectively. •',\n",
              "   'The mean of the class-wise Intersection over Union (mIoU), i.e., the average across classes of the Intersection- Over-Union (deﬁned as IOU = T P T P +F P +F N ).',\n",
              "   '2'],\n",
              "  'page_sentence_count_spacy': 36}]"
            ]
          },
          "metadata": {},
          "execution_count": 9
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "df = pd.DataFrame(pages_and_texts)\n",
        "df.describe().round(2)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 320
        },
        "id": "jdSPv557fzfQ",
        "outputId": "26b610b0-1215-49aa-e7b1-59cbb08f6de7"
      },
      "execution_count": 10,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "       page_number  page_char_count  page_word_count  page_sentence_count_raw  \\\n",
              "count        11.00            11.00            11.00                    11.00   \n",
              "mean        -36.00          3227.00           515.82                    30.55   \n",
              "std           3.32          1376.29           219.43                    25.75   \n",
              "min         -41.00           359.00            59.00                     7.00   \n",
              "25%         -38.50          2469.50           400.50                    16.00   \n",
              "50%         -36.00          3091.00           514.00                    24.00   \n",
              "75%         -33.50          3967.50           636.50                    34.50   \n",
              "max         -31.00          5351.00           814.00                   101.00   \n",
              "\n",
              "       page_token_count  page_sentence_count_spacy  \n",
              "count             11.00                      11.00  \n",
              "mean             806.75                      27.18  \n",
              "std              344.07                      19.65  \n",
              "min               89.75                       5.00  \n",
              "25%              617.38                      15.00  \n",
              "50%              772.75                      24.00  \n",
              "75%              991.88                      34.50  \n",
              "max             1337.75                      75.00  "
            ],
            "text/html": [
              "\n",
              "  <div id=\"df-de19a5be-3b20-4ce6-a9e9-f9e0d6a68594\" class=\"colab-df-container\">\n",
              "    <div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>page_number</th>\n",
              "      <th>page_char_count</th>\n",
              "      <th>page_word_count</th>\n",
              "      <th>page_sentence_count_raw</th>\n",
              "      <th>page_token_count</th>\n",
              "      <th>page_sentence_count_spacy</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>count</th>\n",
              "      <td>11.00</td>\n",
              "      <td>11.00</td>\n",
              "      <td>11.00</td>\n",
              "      <td>11.00</td>\n",
              "      <td>11.00</td>\n",
              "      <td>11.00</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>mean</th>\n",
              "      <td>-36.00</td>\n",
              "      <td>3227.00</td>\n",
              "      <td>515.82</td>\n",
              "      <td>30.55</td>\n",
              "      <td>806.75</td>\n",
              "      <td>27.18</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>std</th>\n",
              "      <td>3.32</td>\n",
              "      <td>1376.29</td>\n",
              "      <td>219.43</td>\n",
              "      <td>25.75</td>\n",
              "      <td>344.07</td>\n",
              "      <td>19.65</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>min</th>\n",
              "      <td>-41.00</td>\n",
              "      <td>359.00</td>\n",
              "      <td>59.00</td>\n",
              "      <td>7.00</td>\n",
              "      <td>89.75</td>\n",
              "      <td>5.00</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>25%</th>\n",
              "      <td>-38.50</td>\n",
              "      <td>2469.50</td>\n",
              "      <td>400.50</td>\n",
              "      <td>16.00</td>\n",
              "      <td>617.38</td>\n",
              "      <td>15.00</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>50%</th>\n",
              "      <td>-36.00</td>\n",
              "      <td>3091.00</td>\n",
              "      <td>514.00</td>\n",
              "      <td>24.00</td>\n",
              "      <td>772.75</td>\n",
              "      <td>24.00</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>75%</th>\n",
              "      <td>-33.50</td>\n",
              "      <td>3967.50</td>\n",
              "      <td>636.50</td>\n",
              "      <td>34.50</td>\n",
              "      <td>991.88</td>\n",
              "      <td>34.50</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>max</th>\n",
              "      <td>-31.00</td>\n",
              "      <td>5351.00</td>\n",
              "      <td>814.00</td>\n",
              "      <td>101.00</td>\n",
              "      <td>1337.75</td>\n",
              "      <td>75.00</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>\n",
              "    <div class=\"colab-df-buttons\">\n",
              "\n",
              "  <div class=\"colab-df-container\">\n",
              "    <button class=\"colab-df-convert\" onclick=\"convertToInteractive('df-de19a5be-3b20-4ce6-a9e9-f9e0d6a68594')\"\n",
              "            title=\"Convert this dataframe to an interactive table.\"\n",
              "            style=\"display:none;\">\n",
              "\n",
              "  <svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\" viewBox=\"0 -960 960 960\">\n",
              "    <path d=\"M120-120v-720h720v720H120Zm60-500h600v-160H180v160Zm220 220h160v-160H400v160Zm0 220h160v-160H400v160ZM180-400h160v-160H180v160Zm440 0h160v-160H620v160ZM180-180h160v-160H180v160Zm440 0h160v-160H620v160Z\"/>\n",
              "  </svg>\n",
              "    </button>\n",
              "\n",
              "  <style>\n",
              "    .colab-df-container {\n",
              "      display:flex;\n",
              "      gap: 12px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert {\n",
              "      background-color: #E8F0FE;\n",
              "      border: none;\n",
              "      border-radius: 50%;\n",
              "      cursor: pointer;\n",
              "      display: none;\n",
              "      fill: #1967D2;\n",
              "      height: 32px;\n",
              "      padding: 0 0 0 0;\n",
              "      width: 32px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert:hover {\n",
              "      background-color: #E2EBFA;\n",
              "      box-shadow: 0px 1px 2px rgba(60, 64, 67, 0.3), 0px 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
              "      fill: #174EA6;\n",
              "    }\n",
              "\n",
              "    .colab-df-buttons div {\n",
              "      margin-bottom: 4px;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert {\n",
              "      background-color: #3B4455;\n",
              "      fill: #D2E3FC;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert:hover {\n",
              "      background-color: #434B5C;\n",
              "      box-shadow: 0px 1px 3px 1px rgba(0, 0, 0, 0.15);\n",
              "      filter: drop-shadow(0px 1px 2px rgba(0, 0, 0, 0.3));\n",
              "      fill: #FFFFFF;\n",
              "    }\n",
              "  </style>\n",
              "\n",
              "    <script>\n",
              "      const buttonEl =\n",
              "        document.querySelector('#df-de19a5be-3b20-4ce6-a9e9-f9e0d6a68594 button.colab-df-convert');\n",
              "      buttonEl.style.display =\n",
              "        google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
              "\n",
              "      async function convertToInteractive(key) {\n",
              "        const element = document.querySelector('#df-de19a5be-3b20-4ce6-a9e9-f9e0d6a68594');\n",
              "        const dataTable =\n",
              "          await google.colab.kernel.invokeFunction('convertToInteractive',\n",
              "                                                    [key], {});\n",
              "        if (!dataTable) return;\n",
              "\n",
              "        const docLinkHtml = 'Like what you see? Visit the ' +\n",
              "          '<a target=\"_blank\" href=https://colab.research.google.com/notebooks/data_table.ipynb>data table notebook</a>'\n",
              "          + ' to learn more about interactive tables.';\n",
              "        element.innerHTML = '';\n",
              "        dataTable['output_type'] = 'display_data';\n",
              "        await google.colab.output.renderOutput(dataTable, element);\n",
              "        const docLink = document.createElement('div');\n",
              "        docLink.innerHTML = docLinkHtml;\n",
              "        element.appendChild(docLink);\n",
              "      }\n",
              "    </script>\n",
              "  </div>\n",
              "\n",
              "\n",
              "<div id=\"df-0fcffe95-88c8-496a-9585-b6e0ce11ccfe\">\n",
              "  <button class=\"colab-df-quickchart\" onclick=\"quickchart('df-0fcffe95-88c8-496a-9585-b6e0ce11ccfe')\"\n",
              "            title=\"Suggest charts\"\n",
              "            style=\"display:none;\">\n",
              "\n",
              "<svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\"viewBox=\"0 0 24 24\"\n",
              "     width=\"24px\">\n",
              "    <g>\n",
              "        <path d=\"M19 3H5c-1.1 0-2 .9-2 2v14c0 1.1.9 2 2 2h14c1.1 0 2-.9 2-2V5c0-1.1-.9-2-2-2zM9 17H7v-7h2v7zm4 0h-2V7h2v10zm4 0h-2v-4h2v4z\"/>\n",
              "    </g>\n",
              "</svg>\n",
              "  </button>\n",
              "\n",
              "<style>\n",
              "  .colab-df-quickchart {\n",
              "      --bg-color: #E8F0FE;\n",
              "      --fill-color: #1967D2;\n",
              "      --hover-bg-color: #E2EBFA;\n",
              "      --hover-fill-color: #174EA6;\n",
              "      --disabled-fill-color: #AAA;\n",
              "      --disabled-bg-color: #DDD;\n",
              "  }\n",
              "\n",
              "  [theme=dark] .colab-df-quickchart {\n",
              "      --bg-color: #3B4455;\n",
              "      --fill-color: #D2E3FC;\n",
              "      --hover-bg-color: #434B5C;\n",
              "      --hover-fill-color: #FFFFFF;\n",
              "      --disabled-bg-color: #3B4455;\n",
              "      --disabled-fill-color: #666;\n",
              "  }\n",
              "\n",
              "  .colab-df-quickchart {\n",
              "    background-color: var(--bg-color);\n",
              "    border: none;\n",
              "    border-radius: 50%;\n",
              "    cursor: pointer;\n",
              "    display: none;\n",
              "    fill: var(--fill-color);\n",
              "    height: 32px;\n",
              "    padding: 0;\n",
              "    width: 32px;\n",
              "  }\n",
              "\n",
              "  .colab-df-quickchart:hover {\n",
              "    background-color: var(--hover-bg-color);\n",
              "    box-shadow: 0 1px 2px rgba(60, 64, 67, 0.3), 0 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
              "    fill: var(--button-hover-fill-color);\n",
              "  }\n",
              "\n",
              "  .colab-df-quickchart-complete:disabled,\n",
              "  .colab-df-quickchart-complete:disabled:hover {\n",
              "    background-color: var(--disabled-bg-color);\n",
              "    fill: var(--disabled-fill-color);\n",
              "    box-shadow: none;\n",
              "  }\n",
              "\n",
              "  .colab-df-spinner {\n",
              "    border: 2px solid var(--fill-color);\n",
              "    border-color: transparent;\n",
              "    border-bottom-color: var(--fill-color);\n",
              "    animation:\n",
              "      spin 1s steps(1) infinite;\n",
              "  }\n",
              "\n",
              "  @keyframes spin {\n",
              "    0% {\n",
              "      border-color: transparent;\n",
              "      border-bottom-color: var(--fill-color);\n",
              "      border-left-color: var(--fill-color);\n",
              "    }\n",
              "    20% {\n",
              "      border-color: transparent;\n",
              "      border-left-color: var(--fill-color);\n",
              "      border-top-color: var(--fill-color);\n",
              "    }\n",
              "    30% {\n",
              "      border-color: transparent;\n",
              "      border-left-color: var(--fill-color);\n",
              "      border-top-color: var(--fill-color);\n",
              "      border-right-color: var(--fill-color);\n",
              "    }\n",
              "    40% {\n",
              "      border-color: transparent;\n",
              "      border-right-color: var(--fill-color);\n",
              "      border-top-color: var(--fill-color);\n",
              "    }\n",
              "    60% {\n",
              "      border-color: transparent;\n",
              "      border-right-color: var(--fill-color);\n",
              "    }\n",
              "    80% {\n",
              "      border-color: transparent;\n",
              "      border-right-color: var(--fill-color);\n",
              "      border-bottom-color: var(--fill-color);\n",
              "    }\n",
              "    90% {\n",
              "      border-color: transparent;\n",
              "      border-bottom-color: var(--fill-color);\n",
              "    }\n",
              "  }\n",
              "</style>\n",
              "\n",
              "  <script>\n",
              "    async function quickchart(key) {\n",
              "      const quickchartButtonEl =\n",
              "        document.querySelector('#' + key + ' button');\n",
              "      quickchartButtonEl.disabled = true;  // To prevent multiple clicks.\n",
              "      quickchartButtonEl.classList.add('colab-df-spinner');\n",
              "      try {\n",
              "        const charts = await google.colab.kernel.invokeFunction(\n",
              "            'suggestCharts', [key], {});\n",
              "      } catch (error) {\n",
              "        console.error('Error during call to suggestCharts:', error);\n",
              "      }\n",
              "      quickchartButtonEl.classList.remove('colab-df-spinner');\n",
              "      quickchartButtonEl.classList.add('colab-df-quickchart-complete');\n",
              "    }\n",
              "    (() => {\n",
              "      let quickchartButtonEl =\n",
              "        document.querySelector('#df-0fcffe95-88c8-496a-9585-b6e0ce11ccfe button');\n",
              "      quickchartButtonEl.style.display =\n",
              "        google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
              "    })();\n",
              "  </script>\n",
              "</div>\n",
              "\n",
              "    </div>\n",
              "  </div>\n"
            ],
            "application/vnd.google.colaboratory.intrinsic+json": {
              "type": "dataframe",
              "summary": "{\n  \"name\": \"df\",\n  \"rows\": 8,\n  \"fields\": [\n    {\n      \"column\": \"page_number\",\n      \"properties\": {\n        \"dtype\": \"number\",\n        \"std\": 20.30541658910604,\n        \"min\": -41.0,\n        \"max\": 11.0,\n        \"num_unique_values\": 7,\n        \"samples\": [\n          11.0,\n          -36.0,\n          -33.5\n        ],\n        \"semantic_type\": \"\",\n        \"description\": \"\"\n      }\n    },\n    {\n      \"column\": \"page_char_count\",\n      \"properties\": {\n        \"dtype\": \"number\",\n        \"std\": 1820.4982200166878,\n        \"min\": 11.0,\n        \"max\": 5351.0,\n        \"num_unique_values\": 8,\n        \"samples\": [\n          3227.0,\n          3091.0,\n          11.0\n        ],\n        \"semantic_type\": \"\",\n        \"description\": \"\"\n      }\n    },\n    {\n      \"column\": \"page_word_count\",\n      \"properties\": {\n        \"dtype\": \"number\",\n        \"std\": 281.2816749115936,\n        \"min\": 11.0,\n        \"max\": 814.0,\n        \"num_unique_values\": 8,\n        \"samples\": [\n          515.82,\n          514.0,\n          11.0\n        ],\n        \"semantic_type\": \"\",\n        \"description\": \"\"\n      }\n    },\n    {\n      \"column\": \"page_sentence_count_raw\",\n      \"properties\": {\n        \"dtype\": \"number\",\n        \"std\": 29.736005111648744,\n        \"min\": 7.0,\n        \"max\": 101.0,\n        \"num_unique_values\": 8,\n        \"samples\": [\n          30.55,\n          24.0,\n          11.0\n        ],\n        \"semantic_type\": \"\",\n        \"description\": \"\"\n      }\n    },\n    {\n      \"column\": \"page_token_count\",\n      \"properties\": {\n        \"dtype\": \"number\",\n        \"std\": 453.532510661348,\n        \"min\": 11.0,\n        \"max\": 1337.75,\n        \"num_unique_values\": 8,\n        \"samples\": [\n          806.75,\n          772.75,\n          11.0\n        ],\n        \"semantic_type\": \"\",\n        \"description\": \"\"\n      }\n    },\n    {\n      \"column\": \"page_sentence_count_spacy\",\n      \"properties\": {\n        \"dtype\": \"number\",\n        \"std\": 21.727463817219242,\n        \"min\": 5.0,\n        \"max\": 75.0,\n        \"num_unique_values\": 8,\n        \"samples\": [\n          27.18,\n          24.0,\n          11.0\n        ],\n        \"semantic_type\": \"\",\n        \"description\": \"\"\n      }\n    }\n  ]\n}"
            }
          },
          "metadata": {},
          "execution_count": 10
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# Define split size to turn groups of sentences into chunks\n",
        "num_sentence_chunk_size = 10\n",
        "\n",
        "# Create a function that recursively splits a list into desired sizes\n",
        "def split_list(input_list: list,\n",
        "               slice_size: int) -> list[list[str]]:\n",
        "    \"\"\"\n",
        "    Splits the input_list into sublists of size slice_size (or as close as possible).\n",
        "\n",
        "    For example, a list of 17 sentences would be split into two lists of [[10], [7]]\n",
        "    \"\"\"\n",
        "    return [input_list[i:i + slice_size] for i in range(0, len(input_list), slice_size)]\n",
        "\n",
        "# Loop through pages and texts and split sentences into chunks\n",
        "for item in tqdm(pages_and_texts):\n",
        "    item[\"sentence_chunks\"] = split_list(input_list=item[\"sentences\"],\n",
        "                                         slice_size=num_sentence_chunk_size)\n",
        "    item[\"num_chunks\"] = len(item[\"sentence_chunks\"])"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 49,
          "referenced_widgets": [
            "646b7becf4a04502829504d9cfedfb71",
            "4911ab5d742940259ae9c00cb66d9293",
            "3d63ca44ccb84d0dad6824b3d33b2dd6",
            "892ddaf2adae4ea1babdd3795899bf0d",
            "a69da305157f497797161e07fb6f748d",
            "9f8b14798bf846a6a79b06fa69f8fe39",
            "7ae4d744435042189fa96bc81f2cecd8",
            "b86c447201de4da4b71b0c214f3b6a06",
            "3bb772d030a74f28b857b6cf7266986e",
            "401f6e163844443fa8a3135afb15c8df",
            "23d158ac64a74eedad1b217f6147b4a0"
          ]
        },
        "id": "XqAyzjCNf6GC",
        "outputId": "be329931-aa6c-41ad-c486-7c964b5bb047"
      },
      "execution_count": 11,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "  0%|          | 0/11 [00:00<?, ?it/s]"
            ],
            "application/vnd.jupyter.widget-view+json": {
              "version_major": 2,
              "version_minor": 0,
              "model_id": "646b7becf4a04502829504d9cfedfb71"
            }
          },
          "metadata": {}
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# Sample an example from the group (note: many samples have only 1 chunk as they have <=10 sentences total)\n",
        "random.sample(pages_and_texts, k=1)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "wsgPXIFRgA72",
        "outputId": "ed65deb3-4aa1-4414-908c-314ee8ef97ec"
      },
      "execution_count": 12,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "[{'page_number': -39,\n",
              "  'page_char_count': 3091,\n",
              "  'page_word_count': 508,\n",
              "  'page_sentence_count_raw': 24,\n",
              "  'page_token_count': 772.75,\n",
              "  'text': 'Figure 1: An downsampled image from the Cityscapes dataset (left) and the corresponding semantic segmentation target (right), in which the pixels belong to one of the classes {background (blue), road (teal), car (yellow), person (red)}. 3 Baseline model The architecture we use is inspired by a fully convolutional residual network called Efﬁcient Neural Network (ENet) [16]. This network was designed for low latency and minimal resource usage. It is designed as a sequence of blocks, summarized in Table 1. The initial block, shown in the left ﬁgure in Fig. 2, encodes the input into a 32×120×76 tensor, which is then processed by a set of sequential blocks of bottlenecks. The ﬁrst three blocks constitute the downsampling encoder, where each block consists of a series of layers as summarized in the left diagram in Fig. 3. The ﬁnal two blocks provide an upsampling decoder, as illustrated in the right diagram in Fig. 3. The ﬁnal block is shown in the right diagram of Fig. 2. Layer Type Output resolution Initial downsample f0 × 120 × 76 3× bottleneck 1 downsample f1 × 60 × 38 3× bottleneck 2 downsample f2 × 30 × 19 3× bottleneck 3 f3 × 30 × 19 3× bottleneck 4 upsample f4 × 60 × 38 3× bottleneck 5 upsample f5 × 120 × 76 Final upsample 4 × 240 × 152 Table 1: Model architecture parametrized by the number of ﬁlters in the bottlenecks fi, with i = 1, ..., 5. Some differences from the original architecture in [16] is that we do not use asymmetric, dilated, or strided convolutions. To further reduce the resource usage, we use three bottlenecks per block instead of ﬁve, and we merge convolutional layers with batch normalization layers by rescaling convolutional ﬁlter weights with batch normalization parameters (implemented through a QConv2DBatchnorm layer [19]). When we use quantization-aware training, this allows us to directly quantize the merged weights during the forward pass, rather than quantizing the batch normalization parameters and the convolutional ﬁlters separately. This merging of layers saves resources on the FPGA, since only the merged weights are used. Performing the merging already during training, ensures that the weights used during training and during inference are quantized the same way. The baseline ENet model is obtained ﬁxing the six f hyperparameters of Table 1 to (32, 64, 64, 64, 128, 48). This choice results in an architecture with 1.1 · 106 parameters, yielding a mIoU=63.2% and an accuracy of 91.5%. 4 Model compression We consider two compression techniques for the model at hand: ﬁlter-wise homogeneous pruning, obtained by reducing the number of ﬁlters on all the convolutional layers; and quantization, i.e., reducing the number of bits allocated for the numerical representation of the network components and the output of each layer computation. In addition, we use the AutoQKeras library [12], distributed with QKeras, to optimize the numerical representation of each component at training time as a hyperparameter. This is done using a mathematical model of the inference power consumption as a constraint in the loss function. 3',\n",
              "  'sentences': ['Figure 1: An downsampled image from the Cityscapes dataset (left) and the corresponding semantic segmentation target (right), in which the pixels belong to one of the classes {background (blue), road (teal), car (yellow), person (red)}.',\n",
              "   '3 Baseline model The architecture we use is inspired by a fully convolutional residual network called Efﬁcient Neural Network (ENet) [16].',\n",
              "   'This network was designed for low latency and minimal resource usage.',\n",
              "   'It is designed as a sequence of blocks, summarized in Table 1.',\n",
              "   'The initial block, shown in the left ﬁgure in Fig.',\n",
              "   '2, encodes the input into a 32×120×76 tensor, which is then processed by a set of sequential blocks of bottlenecks.',\n",
              "   'The ﬁrst three blocks constitute the downsampling encoder, where each block consists of a series of layers as summarized in the left diagram in Fig.',\n",
              "   '3.',\n",
              "   'The ﬁnal two blocks provide an upsampling decoder, as illustrated in the right diagram in Fig.',\n",
              "   '3.',\n",
              "   'The ﬁnal block is shown in the right diagram of Fig.',\n",
              "   '2.',\n",
              "   'Layer Type Output resolution Initial downsample f0 × 120 × 76 3× bottleneck 1 downsample f1 × 60 × 38 3× bottleneck 2 downsample f2 × 30 × 19 3× bottleneck 3 f3 × 30 × 19 3× bottleneck 4 upsample f4 × 60 × 38 3× bottleneck 5 upsample f5 × 120 × 76 Final upsample 4 × 240 × 152 Table 1: Model architecture parametrized by the number of ﬁlters in the bottlenecks fi, with i = 1, ..., 5.',\n",
              "   'Some differences from the original architecture in [16] is that we do not use asymmetric, dilated, or strided convolutions.',\n",
              "   'To further reduce the resource usage, we use three bottlenecks per block instead of ﬁve, and we merge convolutional layers with batch normalization layers by rescaling convolutional ﬁlter weights with batch normalization parameters (implemented through a QConv2DBatchnorm layer [19]).',\n",
              "   'When we use quantization-aware training, this allows us to directly quantize the merged weights during the forward pass, rather than quantizing the batch normalization parameters and the convolutional ﬁlters separately.',\n",
              "   'This merging of layers saves resources on the FPGA, since only the merged weights are used.',\n",
              "   'Performing the merging already during training, ensures that the weights used during training and during inference are quantized the same way.',\n",
              "   'The baseline ENet model is obtained ﬁxing the six f hyperparameters of Table 1 to (32, 64, 64, 64, 128, 48).',\n",
              "   'This choice results in an architecture with 1.1 · 106 parameters, yielding a mIoU=63.2% and an accuracy of 91.5%.',\n",
              "   '4 Model compression We consider two compression techniques for the model at hand: ﬁlter-wise homogeneous pruning, obtained by reducing the number of ﬁlters on all the convolutional layers; and quantization, i.e., reducing the number of bits allocated for the numerical representation of the network components and the output of each layer computation.',\n",
              "   'In addition, we use the AutoQKeras library [12], distributed with QKeras, to optimize the numerical representation of each component at training time as a hyperparameter.',\n",
              "   'This is done using a mathematical model of the inference power consumption as a constraint in the loss function.',\n",
              "   '3'],\n",
              "  'page_sentence_count_spacy': 24,\n",
              "  'sentence_chunks': [['Figure 1: An downsampled image from the Cityscapes dataset (left) and the corresponding semantic segmentation target (right), in which the pixels belong to one of the classes {background (blue), road (teal), car (yellow), person (red)}.',\n",
              "    '3 Baseline model The architecture we use is inspired by a fully convolutional residual network called Efﬁcient Neural Network (ENet) [16].',\n",
              "    'This network was designed for low latency and minimal resource usage.',\n",
              "    'It is designed as a sequence of blocks, summarized in Table 1.',\n",
              "    'The initial block, shown in the left ﬁgure in Fig.',\n",
              "    '2, encodes the input into a 32×120×76 tensor, which is then processed by a set of sequential blocks of bottlenecks.',\n",
              "    'The ﬁrst three blocks constitute the downsampling encoder, where each block consists of a series of layers as summarized in the left diagram in Fig.',\n",
              "    '3.',\n",
              "    'The ﬁnal two blocks provide an upsampling decoder, as illustrated in the right diagram in Fig.',\n",
              "    '3.'],\n",
              "   ['The ﬁnal block is shown in the right diagram of Fig.',\n",
              "    '2.',\n",
              "    'Layer Type Output resolution Initial downsample f0 × 120 × 76 3× bottleneck 1 downsample f1 × 60 × 38 3× bottleneck 2 downsample f2 × 30 × 19 3× bottleneck 3 f3 × 30 × 19 3× bottleneck 4 upsample f4 × 60 × 38 3× bottleneck 5 upsample f5 × 120 × 76 Final upsample 4 × 240 × 152 Table 1: Model architecture parametrized by the number of ﬁlters in the bottlenecks fi, with i = 1, ..., 5.',\n",
              "    'Some differences from the original architecture in [16] is that we do not use asymmetric, dilated, or strided convolutions.',\n",
              "    'To further reduce the resource usage, we use three bottlenecks per block instead of ﬁve, and we merge convolutional layers with batch normalization layers by rescaling convolutional ﬁlter weights with batch normalization parameters (implemented through a QConv2DBatchnorm layer [19]).',\n",
              "    'When we use quantization-aware training, this allows us to directly quantize the merged weights during the forward pass, rather than quantizing the batch normalization parameters and the convolutional ﬁlters separately.',\n",
              "    'This merging of layers saves resources on the FPGA, since only the merged weights are used.',\n",
              "    'Performing the merging already during training, ensures that the weights used during training and during inference are quantized the same way.',\n",
              "    'The baseline ENet model is obtained ﬁxing the six f hyperparameters of Table 1 to (32, 64, 64, 64, 128, 48).',\n",
              "    'This choice results in an architecture with 1.1 · 106 parameters, yielding a mIoU=63.2% and an accuracy of 91.5%.'],\n",
              "   ['4 Model compression We consider two compression techniques for the model at hand: ﬁlter-wise homogeneous pruning, obtained by reducing the number of ﬁlters on all the convolutional layers; and quantization, i.e., reducing the number of bits allocated for the numerical representation of the network components and the output of each layer computation.',\n",
              "    'In addition, we use the AutoQKeras library [12], distributed with QKeras, to optimize the numerical representation of each component at training time as a hyperparameter.',\n",
              "    'This is done using a mathematical model of the inference power consumption as a constraint in the loss function.',\n",
              "    '3']],\n",
              "  'num_chunks': 3}]"
            ]
          },
          "metadata": {},
          "execution_count": 12
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# Create a DataFrame to get stats\n",
        "df = pd.DataFrame(pages_and_texts)\n",
        "df.describe().round(2)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 320
        },
        "id": "08J1peRIg5nL",
        "outputId": "8c129570-f7b3-4c62-fbcb-a97f0146754d"
      },
      "execution_count": 13,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "       page_number  page_char_count  page_word_count  page_sentence_count_raw  \\\n",
              "count        11.00            11.00            11.00                    11.00   \n",
              "mean        -36.00          3227.00           515.82                    30.55   \n",
              "std           3.32          1376.29           219.43                    25.75   \n",
              "min         -41.00           359.00            59.00                     7.00   \n",
              "25%         -38.50          2469.50           400.50                    16.00   \n",
              "50%         -36.00          3091.00           514.00                    24.00   \n",
              "75%         -33.50          3967.50           636.50                    34.50   \n",
              "max         -31.00          5351.00           814.00                   101.00   \n",
              "\n",
              "       page_token_count  page_sentence_count_spacy  num_chunks  \n",
              "count             11.00                      11.00       11.00  \n",
              "mean             806.75                      27.18        3.18  \n",
              "std              344.07                      19.65        2.04  \n",
              "min               89.75                       5.00        1.00  \n",
              "25%              617.38                      15.00        2.00  \n",
              "50%              772.75                      24.00        3.00  \n",
              "75%              991.88                      34.50        4.00  \n",
              "max             1337.75                      75.00        8.00  "
            ],
            "text/html": [
              "\n",
              "  <div id=\"df-3e3ffdc2-7ae7-4dbf-8a86-5a4d91f1aa05\" class=\"colab-df-container\">\n",
              "    <div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>page_number</th>\n",
              "      <th>page_char_count</th>\n",
              "      <th>page_word_count</th>\n",
              "      <th>page_sentence_count_raw</th>\n",
              "      <th>page_token_count</th>\n",
              "      <th>page_sentence_count_spacy</th>\n",
              "      <th>num_chunks</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>count</th>\n",
              "      <td>11.00</td>\n",
              "      <td>11.00</td>\n",
              "      <td>11.00</td>\n",
              "      <td>11.00</td>\n",
              "      <td>11.00</td>\n",
              "      <td>11.00</td>\n",
              "      <td>11.00</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>mean</th>\n",
              "      <td>-36.00</td>\n",
              "      <td>3227.00</td>\n",
              "      <td>515.82</td>\n",
              "      <td>30.55</td>\n",
              "      <td>806.75</td>\n",
              "      <td>27.18</td>\n",
              "      <td>3.18</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>std</th>\n",
              "      <td>3.32</td>\n",
              "      <td>1376.29</td>\n",
              "      <td>219.43</td>\n",
              "      <td>25.75</td>\n",
              "      <td>344.07</td>\n",
              "      <td>19.65</td>\n",
              "      <td>2.04</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>min</th>\n",
              "      <td>-41.00</td>\n",
              "      <td>359.00</td>\n",
              "      <td>59.00</td>\n",
              "      <td>7.00</td>\n",
              "      <td>89.75</td>\n",
              "      <td>5.00</td>\n",
              "      <td>1.00</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>25%</th>\n",
              "      <td>-38.50</td>\n",
              "      <td>2469.50</td>\n",
              "      <td>400.50</td>\n",
              "      <td>16.00</td>\n",
              "      <td>617.38</td>\n",
              "      <td>15.00</td>\n",
              "      <td>2.00</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>50%</th>\n",
              "      <td>-36.00</td>\n",
              "      <td>3091.00</td>\n",
              "      <td>514.00</td>\n",
              "      <td>24.00</td>\n",
              "      <td>772.75</td>\n",
              "      <td>24.00</td>\n",
              "      <td>3.00</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>75%</th>\n",
              "      <td>-33.50</td>\n",
              "      <td>3967.50</td>\n",
              "      <td>636.50</td>\n",
              "      <td>34.50</td>\n",
              "      <td>991.88</td>\n",
              "      <td>34.50</td>\n",
              "      <td>4.00</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>max</th>\n",
              "      <td>-31.00</td>\n",
              "      <td>5351.00</td>\n",
              "      <td>814.00</td>\n",
              "      <td>101.00</td>\n",
              "      <td>1337.75</td>\n",
              "      <td>75.00</td>\n",
              "      <td>8.00</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>\n",
              "    <div class=\"colab-df-buttons\">\n",
              "\n",
              "  <div class=\"colab-df-container\">\n",
              "    <button class=\"colab-df-convert\" onclick=\"convertToInteractive('df-3e3ffdc2-7ae7-4dbf-8a86-5a4d91f1aa05')\"\n",
              "            title=\"Convert this dataframe to an interactive table.\"\n",
              "            style=\"display:none;\">\n",
              "\n",
              "  <svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\" viewBox=\"0 -960 960 960\">\n",
              "    <path d=\"M120-120v-720h720v720H120Zm60-500h600v-160H180v160Zm220 220h160v-160H400v160Zm0 220h160v-160H400v160ZM180-400h160v-160H180v160Zm440 0h160v-160H620v160ZM180-180h160v-160H180v160Zm440 0h160v-160H620v160Z\"/>\n",
              "  </svg>\n",
              "    </button>\n",
              "\n",
              "  <style>\n",
              "    .colab-df-container {\n",
              "      display:flex;\n",
              "      gap: 12px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert {\n",
              "      background-color: #E8F0FE;\n",
              "      border: none;\n",
              "      border-radius: 50%;\n",
              "      cursor: pointer;\n",
              "      display: none;\n",
              "      fill: #1967D2;\n",
              "      height: 32px;\n",
              "      padding: 0 0 0 0;\n",
              "      width: 32px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert:hover {\n",
              "      background-color: #E2EBFA;\n",
              "      box-shadow: 0px 1px 2px rgba(60, 64, 67, 0.3), 0px 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
              "      fill: #174EA6;\n",
              "    }\n",
              "\n",
              "    .colab-df-buttons div {\n",
              "      margin-bottom: 4px;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert {\n",
              "      background-color: #3B4455;\n",
              "      fill: #D2E3FC;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert:hover {\n",
              "      background-color: #434B5C;\n",
              "      box-shadow: 0px 1px 3px 1px rgba(0, 0, 0, 0.15);\n",
              "      filter: drop-shadow(0px 1px 2px rgba(0, 0, 0, 0.3));\n",
              "      fill: #FFFFFF;\n",
              "    }\n",
              "  </style>\n",
              "\n",
              "    <script>\n",
              "      const buttonEl =\n",
              "        document.querySelector('#df-3e3ffdc2-7ae7-4dbf-8a86-5a4d91f1aa05 button.colab-df-convert');\n",
              "      buttonEl.style.display =\n",
              "        google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
              "\n",
              "      async function convertToInteractive(key) {\n",
              "        const element = document.querySelector('#df-3e3ffdc2-7ae7-4dbf-8a86-5a4d91f1aa05');\n",
              "        const dataTable =\n",
              "          await google.colab.kernel.invokeFunction('convertToInteractive',\n",
              "                                                    [key], {});\n",
              "        if (!dataTable) return;\n",
              "\n",
              "        const docLinkHtml = 'Like what you see? Visit the ' +\n",
              "          '<a target=\"_blank\" href=https://colab.research.google.com/notebooks/data_table.ipynb>data table notebook</a>'\n",
              "          + ' to learn more about interactive tables.';\n",
              "        element.innerHTML = '';\n",
              "        dataTable['output_type'] = 'display_data';\n",
              "        await google.colab.output.renderOutput(dataTable, element);\n",
              "        const docLink = document.createElement('div');\n",
              "        docLink.innerHTML = docLinkHtml;\n",
              "        element.appendChild(docLink);\n",
              "      }\n",
              "    </script>\n",
              "  </div>\n",
              "\n",
              "\n",
              "<div id=\"df-804e4ace-e995-4dc1-bd5f-5f175a6ff516\">\n",
              "  <button class=\"colab-df-quickchart\" onclick=\"quickchart('df-804e4ace-e995-4dc1-bd5f-5f175a6ff516')\"\n",
              "            title=\"Suggest charts\"\n",
              "            style=\"display:none;\">\n",
              "\n",
              "<svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\"viewBox=\"0 0 24 24\"\n",
              "     width=\"24px\">\n",
              "    <g>\n",
              "        <path d=\"M19 3H5c-1.1 0-2 .9-2 2v14c0 1.1.9 2 2 2h14c1.1 0 2-.9 2-2V5c0-1.1-.9-2-2-2zM9 17H7v-7h2v7zm4 0h-2V7h2v10zm4 0h-2v-4h2v4z\"/>\n",
              "    </g>\n",
              "</svg>\n",
              "  </button>\n",
              "\n",
              "<style>\n",
              "  .colab-df-quickchart {\n",
              "      --bg-color: #E8F0FE;\n",
              "      --fill-color: #1967D2;\n",
              "      --hover-bg-color: #E2EBFA;\n",
              "      --hover-fill-color: #174EA6;\n",
              "      --disabled-fill-color: #AAA;\n",
              "      --disabled-bg-color: #DDD;\n",
              "  }\n",
              "\n",
              "  [theme=dark] .colab-df-quickchart {\n",
              "      --bg-color: #3B4455;\n",
              "      --fill-color: #D2E3FC;\n",
              "      --hover-bg-color: #434B5C;\n",
              "      --hover-fill-color: #FFFFFF;\n",
              "      --disabled-bg-color: #3B4455;\n",
              "      --disabled-fill-color: #666;\n",
              "  }\n",
              "\n",
              "  .colab-df-quickchart {\n",
              "    background-color: var(--bg-color);\n",
              "    border: none;\n",
              "    border-radius: 50%;\n",
              "    cursor: pointer;\n",
              "    display: none;\n",
              "    fill: var(--fill-color);\n",
              "    height: 32px;\n",
              "    padding: 0;\n",
              "    width: 32px;\n",
              "  }\n",
              "\n",
              "  .colab-df-quickchart:hover {\n",
              "    background-color: var(--hover-bg-color);\n",
              "    box-shadow: 0 1px 2px rgba(60, 64, 67, 0.3), 0 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
              "    fill: var(--button-hover-fill-color);\n",
              "  }\n",
              "\n",
              "  .colab-df-quickchart-complete:disabled,\n",
              "  .colab-df-quickchart-complete:disabled:hover {\n",
              "    background-color: var(--disabled-bg-color);\n",
              "    fill: var(--disabled-fill-color);\n",
              "    box-shadow: none;\n",
              "  }\n",
              "\n",
              "  .colab-df-spinner {\n",
              "    border: 2px solid var(--fill-color);\n",
              "    border-color: transparent;\n",
              "    border-bottom-color: var(--fill-color);\n",
              "    animation:\n",
              "      spin 1s steps(1) infinite;\n",
              "  }\n",
              "\n",
              "  @keyframes spin {\n",
              "    0% {\n",
              "      border-color: transparent;\n",
              "      border-bottom-color: var(--fill-color);\n",
              "      border-left-color: var(--fill-color);\n",
              "    }\n",
              "    20% {\n",
              "      border-color: transparent;\n",
              "      border-left-color: var(--fill-color);\n",
              "      border-top-color: var(--fill-color);\n",
              "    }\n",
              "    30% {\n",
              "      border-color: transparent;\n",
              "      border-left-color: var(--fill-color);\n",
              "      border-top-color: var(--fill-color);\n",
              "      border-right-color: var(--fill-color);\n",
              "    }\n",
              "    40% {\n",
              "      border-color: transparent;\n",
              "      border-right-color: var(--fill-color);\n",
              "      border-top-color: var(--fill-color);\n",
              "    }\n",
              "    60% {\n",
              "      border-color: transparent;\n",
              "      border-right-color: var(--fill-color);\n",
              "    }\n",
              "    80% {\n",
              "      border-color: transparent;\n",
              "      border-right-color: var(--fill-color);\n",
              "      border-bottom-color: var(--fill-color);\n",
              "    }\n",
              "    90% {\n",
              "      border-color: transparent;\n",
              "      border-bottom-color: var(--fill-color);\n",
              "    }\n",
              "  }\n",
              "</style>\n",
              "\n",
              "  <script>\n",
              "    async function quickchart(key) {\n",
              "      const quickchartButtonEl =\n",
              "        document.querySelector('#' + key + ' button');\n",
              "      quickchartButtonEl.disabled = true;  // To prevent multiple clicks.\n",
              "      quickchartButtonEl.classList.add('colab-df-spinner');\n",
              "      try {\n",
              "        const charts = await google.colab.kernel.invokeFunction(\n",
              "            'suggestCharts', [key], {});\n",
              "      } catch (error) {\n",
              "        console.error('Error during call to suggestCharts:', error);\n",
              "      }\n",
              "      quickchartButtonEl.classList.remove('colab-df-spinner');\n",
              "      quickchartButtonEl.classList.add('colab-df-quickchart-complete');\n",
              "    }\n",
              "    (() => {\n",
              "      let quickchartButtonEl =\n",
              "        document.querySelector('#df-804e4ace-e995-4dc1-bd5f-5f175a6ff516 button');\n",
              "      quickchartButtonEl.style.display =\n",
              "        google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
              "    })();\n",
              "  </script>\n",
              "</div>\n",
              "\n",
              "    </div>\n",
              "  </div>\n"
            ],
            "application/vnd.google.colaboratory.intrinsic+json": {
              "type": "dataframe",
              "summary": "{\n  \"name\": \"df\",\n  \"rows\": 8,\n  \"fields\": [\n    {\n      \"column\": \"page_number\",\n      \"properties\": {\n        \"dtype\": \"number\",\n        \"std\": 20.30541658910604,\n        \"min\": -41.0,\n        \"max\": 11.0,\n        \"num_unique_values\": 7,\n        \"samples\": [\n          11.0,\n          -36.0,\n          -33.5\n        ],\n        \"semantic_type\": \"\",\n        \"description\": \"\"\n      }\n    },\n    {\n      \"column\": \"page_char_count\",\n      \"properties\": {\n        \"dtype\": \"number\",\n        \"std\": 1820.4982200166878,\n        \"min\": 11.0,\n        \"max\": 5351.0,\n        \"num_unique_values\": 8,\n        \"samples\": [\n          3227.0,\n          3091.0,\n          11.0\n        ],\n        \"semantic_type\": \"\",\n        \"description\": \"\"\n      }\n    },\n    {\n      \"column\": \"page_word_count\",\n      \"properties\": {\n        \"dtype\": \"number\",\n        \"std\": 281.2816749115936,\n        \"min\": 11.0,\n        \"max\": 814.0,\n        \"num_unique_values\": 8,\n        \"samples\": [\n          515.82,\n          514.0,\n          11.0\n        ],\n        \"semantic_type\": \"\",\n        \"description\": \"\"\n      }\n    },\n    {\n      \"column\": \"page_sentence_count_raw\",\n      \"properties\": {\n        \"dtype\": \"number\",\n        \"std\": 29.736005111648744,\n        \"min\": 7.0,\n        \"max\": 101.0,\n        \"num_unique_values\": 8,\n        \"samples\": [\n          30.55,\n          24.0,\n          11.0\n        ],\n        \"semantic_type\": \"\",\n        \"description\": \"\"\n      }\n    },\n    {\n      \"column\": \"page_token_count\",\n      \"properties\": {\n        \"dtype\": \"number\",\n        \"std\": 453.532510661348,\n        \"min\": 11.0,\n        \"max\": 1337.75,\n        \"num_unique_values\": 8,\n        \"samples\": [\n          806.75,\n          772.75,\n          11.0\n        ],\n        \"semantic_type\": \"\",\n        \"description\": \"\"\n      }\n    },\n    {\n      \"column\": \"page_sentence_count_spacy\",\n      \"properties\": {\n        \"dtype\": \"number\",\n        \"std\": 21.727463817219242,\n        \"min\": 5.0,\n        \"max\": 75.0,\n        \"num_unique_values\": 8,\n        \"samples\": [\n          27.18,\n          24.0,\n          11.0\n        ],\n        \"semantic_type\": \"\",\n        \"description\": \"\"\n      }\n    },\n    {\n      \"column\": \"num_chunks\",\n      \"properties\": {\n        \"dtype\": \"number\",\n        \"std\": 3.441302701843342,\n        \"min\": 1.0,\n        \"max\": 11.0,\n        \"num_unique_values\": 8,\n        \"samples\": [\n          3.18,\n          3.0,\n          11.0\n        ],\n        \"semantic_type\": \"\",\n        \"description\": \"\"\n      }\n    }\n  ]\n}"
            }
          },
          "metadata": {},
          "execution_count": 13
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "import re\n",
        "\n",
        "# Split each chunk into its own item\n",
        "pages_and_chunks = []\n",
        "for item in tqdm(pages_and_texts):\n",
        "    for sentence_chunk in item[\"sentence_chunks\"]:\n",
        "        chunk_dict = {}\n",
        "        chunk_dict[\"page_number\"] = item[\"page_number\"]\n",
        "\n",
        "        # Join the sentences together into a paragraph-like structure, aka a chunk (so they are a single string)\n",
        "        joined_sentence_chunk = \"\".join(sentence_chunk).replace(\"  \", \" \").strip()\n",
        "        joined_sentence_chunk = re.sub(r'\\.([A-Z])', r'. \\1', joined_sentence_chunk) # \".A\" -> \". A\" for any full-stop/capital letter combo\n",
        "        chunk_dict[\"sentence_chunk\"] = joined_sentence_chunk\n",
        "\n",
        "        # Get stats about the chunk\n",
        "        chunk_dict[\"chunk_char_count\"] = len(joined_sentence_chunk)\n",
        "        chunk_dict[\"chunk_word_count\"] = len([word for word in joined_sentence_chunk.split(\" \")])\n",
        "        chunk_dict[\"chunk_token_count\"] = len(joined_sentence_chunk) / 4 # 1 token = ~4 characters\n",
        "\n",
        "        pages_and_chunks.append(chunk_dict)\n",
        "\n",
        "# How many chunks do we have?\n",
        "len(pages_and_chunks)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 67,
          "referenced_widgets": [
            "847db1d1903d4543b1d8532629158bc6",
            "7163d2588d0b47d8809c200e34bc27f4",
            "e1641ed76f1b45659125c3a96774bac9",
            "976f66fb48f647a1b0941b98bb341876",
            "23e1e281c9864a85abe5d8bbf24dc859",
            "988c783606c740339c09bfe5d126f782",
            "49d8c16e33af4224b30d2caccc66c140",
            "66a35f25189447d48ce4b3506488c35a",
            "7a6462ef583a4d65a0e8cae6c6962bfa",
            "65cd40e423df4f69989385b92549f3dc",
            "08116949b49645afb5d3a1d8232f6139"
          ]
        },
        "id": "_8sdkTmphA2J",
        "outputId": "128d6882-fed5-4bc9-e1c8-f5e55d8ebf21"
      },
      "execution_count": 14,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "  0%|          | 0/11 [00:00<?, ?it/s]"
            ],
            "application/vnd.jupyter.widget-view+json": {
              "version_major": 2,
              "version_minor": 0,
              "model_id": "847db1d1903d4543b1d8532629158bc6"
            }
          },
          "metadata": {}
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "35"
            ]
          },
          "metadata": {},
          "execution_count": 14
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# View a random sample\n",
        "random.sample(pages_and_chunks, k=1)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "zbI81I3WhGQ0",
        "outputId": "e11ae681-8a7f-42cb-da33-e506b14c05bb"
      },
      "execution_count": 15,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "[{'page_number': -31,\n",
              "  'sentence_chunk': '8, 675–686. [23] W. Jia, J. Cui, X. Zheng, and Q. Wu, “Design and implementation of real-time semantic segmentation network based on fpga”, in 2021 7th International Conference on Computing and Artiﬁcial Intelligence, ICCAI 2021, p. 321–325. Association for Computing Machinery, New York, NY, USA, 2021.doi:10.1145/3467707.3467756.11',\n",
              "  'chunk_char_count': 333,\n",
              "  'chunk_word_count': 46,\n",
              "  'chunk_token_count': 83.25}]"
            ]
          },
          "metadata": {},
          "execution_count": 15
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# Get stats about our chunks\n",
        "df = pd.DataFrame(pages_and_chunks)\n",
        "df.describe().round(2)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 300
        },
        "id": "25RnEoiohMhK",
        "outputId": "81eaa579-a226-4d7c-bb29-3cb6732ed077"
      },
      "execution_count": 16,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "       page_number  chunk_char_count  chunk_word_count  chunk_token_count\n",
              "count        35.00             35.00             35.00              35.00\n",
              "mean        -35.03           1012.26            160.86             253.06\n",
              "std           3.31            590.47             99.06             147.62\n",
              "min         -41.00             66.00             10.00              16.50\n",
              "25%         -38.00            491.50             66.00             122.88\n",
              "50%         -34.00            992.00            164.00             248.00\n",
              "75%         -32.50           1432.00            231.50             358.00\n",
              "max         -31.00           2364.00            425.00             591.00"
            ],
            "text/html": [
              "\n",
              "  <div id=\"df-3456832b-c31b-4631-b0a1-196b87be6f5e\" class=\"colab-df-container\">\n",
              "    <div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>page_number</th>\n",
              "      <th>chunk_char_count</th>\n",
              "      <th>chunk_word_count</th>\n",
              "      <th>chunk_token_count</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>count</th>\n",
              "      <td>35.00</td>\n",
              "      <td>35.00</td>\n",
              "      <td>35.00</td>\n",
              "      <td>35.00</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>mean</th>\n",
              "      <td>-35.03</td>\n",
              "      <td>1012.26</td>\n",
              "      <td>160.86</td>\n",
              "      <td>253.06</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>std</th>\n",
              "      <td>3.31</td>\n",
              "      <td>590.47</td>\n",
              "      <td>99.06</td>\n",
              "      <td>147.62</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>min</th>\n",
              "      <td>-41.00</td>\n",
              "      <td>66.00</td>\n",
              "      <td>10.00</td>\n",
              "      <td>16.50</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>25%</th>\n",
              "      <td>-38.00</td>\n",
              "      <td>491.50</td>\n",
              "      <td>66.00</td>\n",
              "      <td>122.88</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>50%</th>\n",
              "      <td>-34.00</td>\n",
              "      <td>992.00</td>\n",
              "      <td>164.00</td>\n",
              "      <td>248.00</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>75%</th>\n",
              "      <td>-32.50</td>\n",
              "      <td>1432.00</td>\n",
              "      <td>231.50</td>\n",
              "      <td>358.00</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>max</th>\n",
              "      <td>-31.00</td>\n",
              "      <td>2364.00</td>\n",
              "      <td>425.00</td>\n",
              "      <td>591.00</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>\n",
              "    <div class=\"colab-df-buttons\">\n",
              "\n",
              "  <div class=\"colab-df-container\">\n",
              "    <button class=\"colab-df-convert\" onclick=\"convertToInteractive('df-3456832b-c31b-4631-b0a1-196b87be6f5e')\"\n",
              "            title=\"Convert this dataframe to an interactive table.\"\n",
              "            style=\"display:none;\">\n",
              "\n",
              "  <svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\" viewBox=\"0 -960 960 960\">\n",
              "    <path d=\"M120-120v-720h720v720H120Zm60-500h600v-160H180v160Zm220 220h160v-160H400v160Zm0 220h160v-160H400v160ZM180-400h160v-160H180v160Zm440 0h160v-160H620v160ZM180-180h160v-160H180v160Zm440 0h160v-160H620v160Z\"/>\n",
              "  </svg>\n",
              "    </button>\n",
              "\n",
              "  <style>\n",
              "    .colab-df-container {\n",
              "      display:flex;\n",
              "      gap: 12px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert {\n",
              "      background-color: #E8F0FE;\n",
              "      border: none;\n",
              "      border-radius: 50%;\n",
              "      cursor: pointer;\n",
              "      display: none;\n",
              "      fill: #1967D2;\n",
              "      height: 32px;\n",
              "      padding: 0 0 0 0;\n",
              "      width: 32px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert:hover {\n",
              "      background-color: #E2EBFA;\n",
              "      box-shadow: 0px 1px 2px rgba(60, 64, 67, 0.3), 0px 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
              "      fill: #174EA6;\n",
              "    }\n",
              "\n",
              "    .colab-df-buttons div {\n",
              "      margin-bottom: 4px;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert {\n",
              "      background-color: #3B4455;\n",
              "      fill: #D2E3FC;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert:hover {\n",
              "      background-color: #434B5C;\n",
              "      box-shadow: 0px 1px 3px 1px rgba(0, 0, 0, 0.15);\n",
              "      filter: drop-shadow(0px 1px 2px rgba(0, 0, 0, 0.3));\n",
              "      fill: #FFFFFF;\n",
              "    }\n",
              "  </style>\n",
              "\n",
              "    <script>\n",
              "      const buttonEl =\n",
              "        document.querySelector('#df-3456832b-c31b-4631-b0a1-196b87be6f5e button.colab-df-convert');\n",
              "      buttonEl.style.display =\n",
              "        google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
              "\n",
              "      async function convertToInteractive(key) {\n",
              "        const element = document.querySelector('#df-3456832b-c31b-4631-b0a1-196b87be6f5e');\n",
              "        const dataTable =\n",
              "          await google.colab.kernel.invokeFunction('convertToInteractive',\n",
              "                                                    [key], {});\n",
              "        if (!dataTable) return;\n",
              "\n",
              "        const docLinkHtml = 'Like what you see? Visit the ' +\n",
              "          '<a target=\"_blank\" href=https://colab.research.google.com/notebooks/data_table.ipynb>data table notebook</a>'\n",
              "          + ' to learn more about interactive tables.';\n",
              "        element.innerHTML = '';\n",
              "        dataTable['output_type'] = 'display_data';\n",
              "        await google.colab.output.renderOutput(dataTable, element);\n",
              "        const docLink = document.createElement('div');\n",
              "        docLink.innerHTML = docLinkHtml;\n",
              "        element.appendChild(docLink);\n",
              "      }\n",
              "    </script>\n",
              "  </div>\n",
              "\n",
              "\n",
              "<div id=\"df-5983f58c-4073-431f-ae9d-5af76ba89253\">\n",
              "  <button class=\"colab-df-quickchart\" onclick=\"quickchart('df-5983f58c-4073-431f-ae9d-5af76ba89253')\"\n",
              "            title=\"Suggest charts\"\n",
              "            style=\"display:none;\">\n",
              "\n",
              "<svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\"viewBox=\"0 0 24 24\"\n",
              "     width=\"24px\">\n",
              "    <g>\n",
              "        <path d=\"M19 3H5c-1.1 0-2 .9-2 2v14c0 1.1.9 2 2 2h14c1.1 0 2-.9 2-2V5c0-1.1-.9-2-2-2zM9 17H7v-7h2v7zm4 0h-2V7h2v10zm4 0h-2v-4h2v4z\"/>\n",
              "    </g>\n",
              "</svg>\n",
              "  </button>\n",
              "\n",
              "<style>\n",
              "  .colab-df-quickchart {\n",
              "      --bg-color: #E8F0FE;\n",
              "      --fill-color: #1967D2;\n",
              "      --hover-bg-color: #E2EBFA;\n",
              "      --hover-fill-color: #174EA6;\n",
              "      --disabled-fill-color: #AAA;\n",
              "      --disabled-bg-color: #DDD;\n",
              "  }\n",
              "\n",
              "  [theme=dark] .colab-df-quickchart {\n",
              "      --bg-color: #3B4455;\n",
              "      --fill-color: #D2E3FC;\n",
              "      --hover-bg-color: #434B5C;\n",
              "      --hover-fill-color: #FFFFFF;\n",
              "      --disabled-bg-color: #3B4455;\n",
              "      --disabled-fill-color: #666;\n",
              "  }\n",
              "\n",
              "  .colab-df-quickchart {\n",
              "    background-color: var(--bg-color);\n",
              "    border: none;\n",
              "    border-radius: 50%;\n",
              "    cursor: pointer;\n",
              "    display: none;\n",
              "    fill: var(--fill-color);\n",
              "    height: 32px;\n",
              "    padding: 0;\n",
              "    width: 32px;\n",
              "  }\n",
              "\n",
              "  .colab-df-quickchart:hover {\n",
              "    background-color: var(--hover-bg-color);\n",
              "    box-shadow: 0 1px 2px rgba(60, 64, 67, 0.3), 0 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
              "    fill: var(--button-hover-fill-color);\n",
              "  }\n",
              "\n",
              "  .colab-df-quickchart-complete:disabled,\n",
              "  .colab-df-quickchart-complete:disabled:hover {\n",
              "    background-color: var(--disabled-bg-color);\n",
              "    fill: var(--disabled-fill-color);\n",
              "    box-shadow: none;\n",
              "  }\n",
              "\n",
              "  .colab-df-spinner {\n",
              "    border: 2px solid var(--fill-color);\n",
              "    border-color: transparent;\n",
              "    border-bottom-color: var(--fill-color);\n",
              "    animation:\n",
              "      spin 1s steps(1) infinite;\n",
              "  }\n",
              "\n",
              "  @keyframes spin {\n",
              "    0% {\n",
              "      border-color: transparent;\n",
              "      border-bottom-color: var(--fill-color);\n",
              "      border-left-color: var(--fill-color);\n",
              "    }\n",
              "    20% {\n",
              "      border-color: transparent;\n",
              "      border-left-color: var(--fill-color);\n",
              "      border-top-color: var(--fill-color);\n",
              "    }\n",
              "    30% {\n",
              "      border-color: transparent;\n",
              "      border-left-color: var(--fill-color);\n",
              "      border-top-color: var(--fill-color);\n",
              "      border-right-color: var(--fill-color);\n",
              "    }\n",
              "    40% {\n",
              "      border-color: transparent;\n",
              "      border-right-color: var(--fill-color);\n",
              "      border-top-color: var(--fill-color);\n",
              "    }\n",
              "    60% {\n",
              "      border-color: transparent;\n",
              "      border-right-color: var(--fill-color);\n",
              "    }\n",
              "    80% {\n",
              "      border-color: transparent;\n",
              "      border-right-color: var(--fill-color);\n",
              "      border-bottom-color: var(--fill-color);\n",
              "    }\n",
              "    90% {\n",
              "      border-color: transparent;\n",
              "      border-bottom-color: var(--fill-color);\n",
              "    }\n",
              "  }\n",
              "</style>\n",
              "\n",
              "  <script>\n",
              "    async function quickchart(key) {\n",
              "      const quickchartButtonEl =\n",
              "        document.querySelector('#' + key + ' button');\n",
              "      quickchartButtonEl.disabled = true;  // To prevent multiple clicks.\n",
              "      quickchartButtonEl.classList.add('colab-df-spinner');\n",
              "      try {\n",
              "        const charts = await google.colab.kernel.invokeFunction(\n",
              "            'suggestCharts', [key], {});\n",
              "      } catch (error) {\n",
              "        console.error('Error during call to suggestCharts:', error);\n",
              "      }\n",
              "      quickchartButtonEl.classList.remove('colab-df-spinner');\n",
              "      quickchartButtonEl.classList.add('colab-df-quickchart-complete');\n",
              "    }\n",
              "    (() => {\n",
              "      let quickchartButtonEl =\n",
              "        document.querySelector('#df-5983f58c-4073-431f-ae9d-5af76ba89253 button');\n",
              "      quickchartButtonEl.style.display =\n",
              "        google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
              "    })();\n",
              "  </script>\n",
              "</div>\n",
              "\n",
              "    </div>\n",
              "  </div>\n"
            ],
            "application/vnd.google.colaboratory.intrinsic+json": {
              "type": "dataframe",
              "summary": "{\n  \"name\": \"df\",\n  \"rows\": 8,\n  \"fields\": [\n    {\n      \"column\": \"page_number\",\n      \"properties\": {\n        \"dtype\": \"number\",\n        \"std\": 26.754601606879838,\n        \"min\": -41.0,\n        \"max\": 35.0,\n        \"num_unique_values\": 8,\n        \"samples\": [\n          -35.03,\n          -34.0,\n          35.0\n        ],\n        \"semantic_type\": \"\",\n        \"description\": \"\"\n      }\n    },\n    {\n      \"column\": \"chunk_char_count\",\n      \"properties\": {\n        \"dtype\": \"number\",\n        \"std\": 769.7842670317723,\n        \"min\": 35.0,\n        \"max\": 2364.0,\n        \"num_unique_values\": 8,\n        \"samples\": [\n          1012.26,\n          992.0,\n          35.0\n        ],\n        \"semantic_type\": \"\",\n        \"description\": \"\"\n      }\n    },\n    {\n      \"column\": \"chunk_word_count\",\n      \"properties\": {\n        \"dtype\": \"number\",\n        \"std\": 133.65620671280905,\n        \"min\": 10.0,\n        \"max\": 425.0,\n        \"num_unique_values\": 8,\n        \"samples\": [\n          160.86,\n          164.0,\n          35.0\n        ],\n        \"semantic_type\": \"\",\n        \"description\": \"\"\n      }\n    },\n    {\n      \"column\": \"chunk_token_count\",\n      \"properties\": {\n        \"dtype\": \"number\",\n        \"std\": 188.54790673914755,\n        \"min\": 16.5,\n        \"max\": 591.0,\n        \"num_unique_values\": 8,\n        \"samples\": [\n          253.06,\n          248.0,\n          35.0\n        ],\n        \"semantic_type\": \"\",\n        \"description\": \"\"\n      }\n    }\n  ]\n}"
            }
          },
          "metadata": {},
          "execution_count": 16
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# Show random chunks with under 30 tokens in length\n",
        "min_token_length = 10\n",
        "for row in df[df[\"chunk_token_count\"] <= min_token_length].sample(3).iterrows():\n",
        "    print(f'Chunk token count: {row[1][\"chunk_token_count\"]} | Text: {row[1][\"sentence_chunk\"]}')"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 393
        },
        "id": "0zH5yoV1hURu",
        "outputId": "f515ef1b-43d5-42ce-e16d-3de156176d5d"
      },
      "execution_count": 20,
      "outputs": [
        {
          "output_type": "error",
          "ename": "ValueError",
          "evalue": "a must be greater than 0 unless no samples are taken",
          "traceback": [
            "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m",
            "\u001b[0;31mValueError\u001b[0m                                Traceback (most recent call last)",
            "\u001b[0;32m<ipython-input-20-fd5375731717>\u001b[0m in \u001b[0;36m<cell line: 3>\u001b[0;34m()\u001b[0m\n\u001b[1;32m      1\u001b[0m \u001b[0;31m# Show random chunks with under 30 tokens in length\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m      2\u001b[0m \u001b[0mmin_token_length\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0;36m10\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m----> 3\u001b[0;31m \u001b[0;32mfor\u001b[0m \u001b[0mrow\u001b[0m \u001b[0;32min\u001b[0m \u001b[0mdf\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0mdf\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;34m\"chunk_token_count\"\u001b[0m\u001b[0;34m]\u001b[0m \u001b[0;34m<=\u001b[0m \u001b[0mmin_token_length\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0msample\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;36m3\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0miterrows\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m      4\u001b[0m     \u001b[0mprint\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34mf'Chunk token count: {row[1][\"chunk_token_count\"]} | Text: {row[1][\"sentence_chunk\"]}'\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
            "\u001b[0;32m/usr/local/lib/python3.10/dist-packages/pandas/core/generic.py\u001b[0m in \u001b[0;36msample\u001b[0;34m(self, n, frac, replace, weights, random_state, axis, ignore_index)\u001b[0m\n\u001b[1;32m   6116\u001b[0m             \u001b[0mweights\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0msample\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mpreprocess_weights\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mweights\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0maxis\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m   6117\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m-> 6118\u001b[0;31m         \u001b[0msampled_indices\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0msample\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0msample\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mobj_len\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0msize\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mreplace\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mweights\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mrs\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m   6119\u001b[0m         \u001b[0mresult\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mtake\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0msampled_indices\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0maxis\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0maxis\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m   6120\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n",
            "\u001b[0;32m/usr/local/lib/python3.10/dist-packages/pandas/core/sample.py\u001b[0m in \u001b[0;36msample\u001b[0;34m(obj_len, size, replace, weights, random_state)\u001b[0m\n\u001b[1;32m    150\u001b[0m             \u001b[0;32mraise\u001b[0m \u001b[0mValueError\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m\"Invalid weights: weights sum to zero\"\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    151\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m--> 152\u001b[0;31m     return random_state.choice(obj_len, size=size, replace=replace, p=weights).astype(\n\u001b[0m\u001b[1;32m    153\u001b[0m         \u001b[0mnp\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mintp\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mcopy\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;32mFalse\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    154\u001b[0m     )\n",
            "\u001b[0;32mnumpy/random/mtrand.pyx\u001b[0m in \u001b[0;36mnumpy.random.mtrand.RandomState.choice\u001b[0;34m()\u001b[0m\n",
            "\u001b[0;31mValueError\u001b[0m: a must be greater than 0 unless no samples are taken"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "pages_and_chunks_over_min_token_len = df[df[\"chunk_token_count\"] > min_token_length].to_dict(orient=\"records\")\n",
        "pages_and_chunks_over_min_token_len[:2]"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "2n1xDQFqh3By",
        "outputId": "596a4ce2-0fcf-41ed-f7fd-db1ecba02119"
      },
      "execution_count": 21,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "[{'page_number': -41,\n",
              "  'sentence_chunk': 'REAL-TIME SEMANTIC SEGMENTATION ON FPGAS FOR AUTONOMOUS VEHICLES WITH HLS4ML Nicolò Ghielmetti∗, Vladimir Loncar†, Maurizio Pierini, Marcel Roed‡, Sioni Summers European Organization for Nuclear Research (CERN) CH-1211 Geneva 23, Switzerland Thea Aarrestad Institute for Particle Physics and Astrophysics, ETH Zürich 8093 Zürich, Switzerland Christoffer Petersson§ Zenseact Gothenburg, 41756, Sweden Hampus Linander University of Gothenburg Gothenburg, 40530, Sweden Jennifer Ngadiuba Fermi National Accelerator Laboratory Batavia, IL 60510, USA Kelvin Lin¶ University of Washington Seattle, WA 98195, USA Philip Harris Massachusetts Institute of Technology Cambridge, MA 02139, USA May 17, 2022 ABSTRACT In this paper, we investigate how ﬁeld programmable gate arrays can serve as hardware accelerators for real-time semantic segmentation tasks relevant for autonomous driving. Considering compressed versions of the ENet convolutional neural network architecture, we demonstrate a fully-on-chip deployment with a latency of 4.9 ms per image, using less than 30% of the available resources on a Xilinx ZCU102 evaluation board. The latency is reduced to 3 ms per image when increasing the batch size to ten, corresponding to the use case where the autonomous vehicle receives inputs from multiple cameras simultaneously. We show, through aggressive ﬁlter reduction and heterogeneous quantization-aware training, and an optimized implementation of convolutional layers, that the power consumption and resource utilization can be signiﬁcantly reduced while maintaining accuracy on the Cityscapes dataset. Keywords deep learning · FPGA · convolutional neural network 1 Introduction Deep Learning has strongly reshaped computer vision in the last decade, bringing the accuracy of image recognition applications to unprecedented levels. Improved pattern recognition capabilities have had a signiﬁcant impact on the advancement of research in science and technology. Many of the challenges faced by future scientiﬁc experiments, such as the CERN High Luminosity LHC [1] or the Square Kilometer Array observatory [2], and technological ∗Also at Politecnico di Milano, Italy †Also at Institute of Physics Belgrade, Serbia ‡Also at University of Oxford, UK §Also at Chalmers University of Technology, Sweden ¶Currently at Amazon, USA arXiv:2205.07690v1 [cs. CV] 16 May 2022',\n",
              "  'chunk_char_count': 2364,\n",
              "  'chunk_word_count': 329,\n",
              "  'chunk_token_count': 591.0},\n",
              " {'page_number': -40,\n",
              "  'sentence_chunk': 'challenges faced by, for example, the automotive industry, will require the capability of processing large amounts of data in real-time, often through edge computing devices with strict latency and power-consumption constraints. This requirement has generated interest in the development of energy-effective neural networks, resulting in efforts like tinyML [3], which aims to reduce power consumption as much as possible without negatively affecting the model accuracy. Advances in Deep Learning for computer vision have had a crucial impact on the development of autonomous vehicles, enabling the vehicles to perceive their environment at ever-increasing levels of accuracy and detail. Deep Neural Networks are used for ﬁnding patterns and extracting relevant information from camera images, such as the precise location of the surrounding vehicles and pedestrians. In order for an autonomous vehicle to drive safely and efﬁciently, it must be able to react fast and make quick decisions. This imposes strict latency requirements on the neural networks that are deployed to run inference on resource-limited embedded hardware in the vehicle. In addition to algorithmic development, computer vision for autonomous vehicles has beneﬁted from technological advances in parallel computing architecture [4]. The possibility of performing network training and inference on graphics processing units (GPUs) has made large and complex networks computationally affordable and testable on real-life problems. Due to their high efﬁciency, GPUs have become a common hardware choice in the automotive industry for on-vehicle Deep Learning inference. In this paper, we investigate the possibility of exploiting ﬁeld-programmable gate arrays (FPGAs) as a low-power, inference-optimized, highly parallelisable alternative to GPUs.',\n",
              "  'chunk_char_count': 1816,\n",
              "  'chunk_word_count': 256,\n",
              "  'chunk_token_count': 454.0}]"
            ]
          },
          "metadata": {},
          "execution_count": 21
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# Requires !pip install sentence-transformers\n",
        "from sentence_transformers import SentenceTransformer\n",
        "embedding_model = SentenceTransformer(model_name_or_path=\"all-mpnet-base-v2\",\n",
        "                                      device=\"cpu\") # choose the device to load the model to (note: GPU will often be *much* faster than CPU)\n",
        "\n",
        "# Create a list of sentences to turn into numbers\n",
        "sentences = [\n",
        "    \"The Sentences Transformers library provides an easy and open-source way to create embeddings.\",\n",
        "    \"Sentences can be embedded one by one or as a list of strings.\",\n",
        "    \"Embeddings are one of the most powerful concepts in machine learning!\",\n",
        "    \"Learn to use embeddings well and you'll be well on your way to being an AI engineer.\"\n",
        "]\n",
        "\n",
        "# Sentences are encoded/embedded by calling model.encode()\n",
        "embeddings = embedding_model.encode(sentences)\n",
        "embeddings_dict = dict(zip(sentences, embeddings))\n",
        "\n",
        "# See the embeddings\n",
        "for sentence, embedding in embeddings_dict.items():\n",
        "    print(\"Sentence:\", sentence)\n",
        "    print(\"Embedding:\", embedding)\n",
        "    print(\"\")"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 1000,
          "referenced_widgets": [
            "917cf9de71c446d1904037d3ff10598c",
            "a0f0732d553042e0a3be3ced90546a1e",
            "c5bd847c659149dabfc2e1ccf99b90e4",
            "73fe14a93678443e9b3290a07e88f6a1",
            "4a1b367d9ccd41ca983ef892198b86b9",
            "25b61ae5d8954d5a947963efae7fa41f",
            "913af069fbb94c1da12f72104712121e",
            "b57d5ad3287b4bb8b73a21e7038be814",
            "5fe4516b0613424fbcbf453a7d430a2b",
            "2d1c40e0bfe346a9891a8583c9ef0087",
            "3bc7dacd564548d181cec9088079fc73",
            "18b4d737f97945f5b56c109158d2d8cf",
            "9b715eb261574cc6ae286b3ba4a5ce5e",
            "d88544be1bc5456691521c725a49c87f",
            "a02ea3cc6ad645329fb54fead34c7533",
            "c8be9055dcaa40e4b78d17362ec00632",
            "01fa3a36918d4a2083ebfc5830a63da0",
            "0b9a20b821d84a218dcd7e7ed7d70d65",
            "ae2fbd69796f40eb8527eb3c08258ec6",
            "550768688aa34f06a3e0f13c1dba0e3b",
            "0eea9da5905149c791d9ff82eea946f1",
            "41ebc1df3f4a47a7a1199078c512e129",
            "c7d41b79979d4e52bebaa1439b085ee8",
            "9a6d1bfea04646f4adc753bc9e4559f8",
            "09201e30592e448690b28882ee985c6b",
            "1e6f7d1986a84bdcbbb621aa7bbd7305",
            "6e5cdca32c414328847bd9aca70f273b",
            "b588003d48d04a199546728f4cc0fde6",
            "ea16b1b46032464fbcd82478ddd6a40c",
            "5f6222157d8042b4b88e9f5178653512",
            "219462e1e94c4b7789412ab4f2b8d50b",
            "0f2c243303fd4c31befbcbf9a2703953",
            "55ec1bcc7ecf4fe88a13da12c71e521b",
            "edb7e6d6a8c9418baf3cc538bc5084b1",
            "00382f4dee2744c1906bc09de6763787",
            "5bd202d2b9e043b79a171860d3fb45d8",
            "d178e61f3ac64c9eb3f27af5f12fce71",
            "509a93c7ab264dcfa622cd53765cbaa7",
            "5293267b40eb4f6fb633bcd410e93b22",
            "28925d888cc84557b75aac7079e50fd5",
            "114db12faa874a73b8479b256dfc124d",
            "ec87b3217fec4aa4812bb3c1cc317f8c",
            "2f50c68c88f545f29d756abb36faee2c",
            "f395239fd1c14e2fa52189b243fa1c45",
            "4e66d50820314451be18f0106cf018d9",
            "6f0745faad3344ffb9a7104996bcd85b",
            "c24bcce292a341c2b026e1a990b6c859",
            "f33f9842b7674f5885c137e235904248",
            "ef6a09eec3d245e292108e9c7887c648",
            "04b07550df0e47aeb80a43e4d1f47d7a",
            "ebd38d359a89452b9d2652f90f8d9fef",
            "00c929f0c32640819be5cc03b59a92c0",
            "e59437e8fd6142cda9fb150b1f6fb555",
            "873a408ae76a41bdb4f0879c551bd405",
            "df61d96dbb6949bda5e9a987484b77a7",
            "d578666f07024c29897b773aa8c36471",
            "640ef099b7444df0ba7ddb91c462d151",
            "662ab50b47d54419b781053f5e54b872",
            "7026d50e81de4ccaa6558254ff62abfc",
            "a38ced2468d5405ab47ba2611b39936f",
            "751ae9190e164ee7bfea2b5ffb31ca95",
            "a0d8825e482e4e828c7de1e7ed0e5993",
            "bc8e982fb5a44004be1cc40594558615",
            "b331e927ffce4e1fa329a4dd0b5c4a40",
            "b1554c6cba7243a4863b2faa217270e0",
            "eb530f24bd874030bda74157a8e1c808",
            "8972b726f8384c27b53d10de9161c5c3",
            "4249003994354e7dbc517d09e1ee3882",
            "1ace14b8664c471786656c99a0bed888",
            "0414bb5095b74f98bdafbbb57660707f",
            "32be678b3a18428aad780c63d429c73e",
            "c036f1444e71425c92acabb9325f7a16",
            "33ab65d148fb473e81df708f11eac353",
            "3f2483f387a94d40bdb37caf019da0da",
            "5928d37584d744a986f80c239d161067",
            "fe709112e0404cc3adc8f5e36b4dee6c",
            "b19e6c67575240c4858574bd593c5b13",
            "561f4a0b88a4483ebe99a0272bfe542e",
            "eec84441708343518e35fee9bc4be234",
            "7caa2e0a1ffc44df81804b743615829c",
            "b7a9da9514b941f38b95f959caba67a3",
            "716a5878c7684e23a8ade614b7029313",
            "bff0f96b6765448c8a0fc62b0c7daab7",
            "c8d8373fa5eb4ed59d7833901775d179",
            "8afe34f0fb7a43a6a4357141a20c3297",
            "c93112f1a32840f5a63604bbdfb060c5",
            "342f30219a9f47e4ad3035a8a50a54b0",
            "3d5c8db5bfbe434b85699b68ca7538ab",
            "311e46ee70694985af3a43ddd29b5d48",
            "d7f60361121344e58ab84284553713b3",
            "ea37dd51374449e799db32052166736e",
            "46bd841bfe734478a4013d12784bb95c",
            "2211b11bc7314d279dbec814880195eb",
            "6ef3dfc72fac41019ab3daf266d4605f",
            "964ec1d142a343a6924c805d9fc5b6d1",
            "2222c590e54a4bb6be7c0f08dfc833f1",
            "debc092429e445898728741e66dbae28",
            "403ff93e2f6c41b084337485ef83cfc2",
            "0fb9563c385d40fc9853a72b743121d4",
            "93b1d3bc25b14cc6832ef660b0208579",
            "2d695f86703840059af0f07f22abe0e5",
            "b381d2fb48cb4c7da9770f77a8e25dc6",
            "70416e4ca3e442d7a35c28dc326259c3",
            "2b7c301c186f475bb7ba43df3ac32d1e",
            "7a2905afb99741ffac956a2a6cbb52cd",
            "4bc1838d2bb8439cbe610a3da0b32f01",
            "e389e0e85bf242ca9f687c8f7e47a9ec",
            "974021bd7cf945f08b115f501ce085ee",
            "c1b195650c8b47179bee7dc5337bac2e",
            "7c3849ffcdce40a0baa395d3ea102f03",
            "4fa081728b234291bc569b2df0611996",
            "5cd36aef0a70430d808d413ec3e47c8b",
            "7870b145821b40dc9e9a839748dca05a",
            "efbb30ec5b864c029eb7ddb61389d1b4",
            "21e9be74d6e145efaf45e58cb5709f32",
            "ef48aa6295584710a2a13a6a8fbb276c",
            "60740ca54036463197c5d039f51298d2",
            "d8dca4a7ce8c4f5aa2b0785ab41705db",
            "29ca0ba7f9114016a164ce51196ef061",
            "5513f837752649fea50bd9c0ea209660",
            "c5061f4db4fd4c3790df4d71d2e1ab4b"
          ]
        },
        "id": "B6n7PAWUiG1C",
        "outputId": "d2f20ddc-9aaa-45fa-c575-992790f320b5"
      },
      "execution_count": 22,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "/usr/local/lib/python3.10/dist-packages/huggingface_hub/utils/_auth.py:104: UserWarning: \n",
            "Error while fetching `HF_TOKEN` secret value from your vault: 'Requesting secret HF_TOKEN timed out. Secrets can only be fetched when running from the Colab UI.'.\n",
            "You are not authenticated with the Hugging Face Hub in this notebook.\n",
            "If the error persists, please let us know by opening an issue on GitHub (https://github.com/huggingface/huggingface_hub/issues/new).\n",
            "  warnings.warn(\n"
          ]
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "modules.json:   0%|          | 0.00/349 [00:00<?, ?B/s]"
            ],
            "application/vnd.jupyter.widget-view+json": {
              "version_major": 2,
              "version_minor": 0,
              "model_id": "917cf9de71c446d1904037d3ff10598c"
            }
          },
          "metadata": {}
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "config_sentence_transformers.json:   0%|          | 0.00/116 [00:00<?, ?B/s]"
            ],
            "application/vnd.jupyter.widget-view+json": {
              "version_major": 2,
              "version_minor": 0,
              "model_id": "18b4d737f97945f5b56c109158d2d8cf"
            }
          },
          "metadata": {}
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "README.md:   0%|          | 0.00/10.6k [00:00<?, ?B/s]"
            ],
            "application/vnd.jupyter.widget-view+json": {
              "version_major": 2,
              "version_minor": 0,
              "model_id": "c7d41b79979d4e52bebaa1439b085ee8"
            }
          },
          "metadata": {}
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "sentence_bert_config.json:   0%|          | 0.00/53.0 [00:00<?, ?B/s]"
            ],
            "application/vnd.jupyter.widget-view+json": {
              "version_major": 2,
              "version_minor": 0,
              "model_id": "edb7e6d6a8c9418baf3cc538bc5084b1"
            }
          },
          "metadata": {}
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "config.json:   0%|          | 0.00/571 [00:00<?, ?B/s]"
            ],
            "application/vnd.jupyter.widget-view+json": {
              "version_major": 2,
              "version_minor": 0,
              "model_id": "4e66d50820314451be18f0106cf018d9"
            }
          },
          "metadata": {}
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "model.safetensors:   0%|          | 0.00/438M [00:00<?, ?B/s]"
            ],
            "application/vnd.jupyter.widget-view+json": {
              "version_major": 2,
              "version_minor": 0,
              "model_id": "d578666f07024c29897b773aa8c36471"
            }
          },
          "metadata": {}
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "tokenizer_config.json:   0%|          | 0.00/363 [00:00<?, ?B/s]"
            ],
            "application/vnd.jupyter.widget-view+json": {
              "version_major": 2,
              "version_minor": 0,
              "model_id": "8972b726f8384c27b53d10de9161c5c3"
            }
          },
          "metadata": {}
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "vocab.txt:   0%|          | 0.00/232k [00:00<?, ?B/s]"
            ],
            "application/vnd.jupyter.widget-view+json": {
              "version_major": 2,
              "version_minor": 0,
              "model_id": "561f4a0b88a4483ebe99a0272bfe542e"
            }
          },
          "metadata": {}
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "tokenizer.json:   0%|          | 0.00/466k [00:00<?, ?B/s]"
            ],
            "application/vnd.jupyter.widget-view+json": {
              "version_major": 2,
              "version_minor": 0,
              "model_id": "311e46ee70694985af3a43ddd29b5d48"
            }
          },
          "metadata": {}
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "special_tokens_map.json:   0%|          | 0.00/239 [00:00<?, ?B/s]"
            ],
            "application/vnd.jupyter.widget-view+json": {
              "version_major": 2,
              "version_minor": 0,
              "model_id": "93b1d3bc25b14cc6832ef660b0208579"
            }
          },
          "metadata": {}
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "1_Pooling/config.json:   0%|          | 0.00/190 [00:00<?, ?B/s]"
            ],
            "application/vnd.jupyter.widget-view+json": {
              "version_major": 2,
              "version_minor": 0,
              "model_id": "4fa081728b234291bc569b2df0611996"
            }
          },
          "metadata": {}
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Sentence: The Sentences Transformers library provides an easy and open-source way to create embeddings.\n",
            "Embedding: [-2.07982697e-02  3.03164832e-02 -2.01217849e-02  6.86484650e-02\n",
            " -2.55256221e-02 -8.47686455e-03 -2.07225574e-04 -6.32377118e-02\n",
            "  2.81606894e-02 -3.33353989e-02  3.02633960e-02  5.30721806e-02\n",
            " -5.03527038e-02  2.62288321e-02  3.33313718e-02 -4.51577231e-02\n",
            "  3.63044813e-02 -1.37122418e-03 -1.20171458e-02  1.14947259e-02\n",
            "  5.04510924e-02  4.70856987e-02  2.11913940e-02  5.14606535e-02\n",
            " -2.03746483e-02 -3.58889401e-02 -6.67763175e-04 -2.94393823e-02\n",
            "  4.95859198e-02 -1.05639677e-02 -1.52014112e-02 -1.31758570e-03\n",
            "  4.48197424e-02  1.56023465e-02  8.60379430e-07 -1.21392624e-03\n",
            " -2.37978697e-02 -9.09368275e-04  7.34484056e-03 -2.53933994e-03\n",
            "  5.23370504e-02 -4.68043424e-02  1.66214760e-02  4.71579395e-02\n",
            " -4.15599644e-02  9.01976076e-04  3.60277519e-02  3.42214219e-02\n",
            "  9.68227163e-02  5.94829023e-02 -1.64984372e-02 -3.51249315e-02\n",
            "  5.92516130e-03 -7.07903586e-04 -2.41031442e-02  3.49740870e-02\n",
            " -2.94746887e-02  6.04262855e-03 -9.80652310e-03  2.83218101e-02\n",
            " -1.85375679e-02  3.63212861e-02  1.30292149e-02 -3.71233001e-02\n",
            "  5.27256243e-02 -1.19707026e-02 -7.18083233e-02  1.24431746e-02\n",
            " -6.70563010e-03  7.42154419e-02  1.16356099e-02 -1.74534023e-02\n",
            " -1.82405282e-02 -1.88931134e-02  2.82415152e-02  1.32829510e-02\n",
            " -3.51909734e-02  8.87303147e-04  5.79572394e-02  3.22094113e-02\n",
            " -3.48579930e-03  4.13768739e-02  1.44358072e-02 -3.28044333e-02\n",
            " -9.79085453e-03 -3.16492803e-02  4.23870534e-02 -4.70846333e-02\n",
            " -2.08936650e-02 -1.91249400e-02 -1.22627662e-02  1.01606082e-02\n",
            "  3.91921513e-02 -2.61896141e-02  1.09027857e-02  1.35722943e-02\n",
            " -5.79267405e-02 -3.21500264e-02 -5.75725595e-03 -2.43515857e-02\n",
            "  5.23416735e-02  5.46121923e-03 -2.30996069e-02  2.57174857e-03\n",
            " -6.63347095e-02  3.54125462e-02 -1.03907324e-02  2.25410350e-02\n",
            " -1.84573866e-02 -2.42007170e-02 -4.78365757e-02 -4.79222648e-03\n",
            " -5.34138344e-02  3.01790610e-02 -1.56131107e-02 -5.51476665e-02\n",
            " -3.91874090e-02  5.92152663e-02 -3.47646885e-02  9.68123414e-03\n",
            "  2.13415958e-02  2.30417158e-02  1.91712864e-02  2.77378634e-02\n",
            " -7.73504097e-03  1.04445815e-02 -2.67719701e-02 -2.40199845e-02\n",
            " -1.92290191e-02  3.91504029e-03 -2.54714545e-02  3.61943170e-02\n",
            "  5.12867272e-02 -8.41695443e-03 -3.13829891e-02  1.47483395e-02\n",
            "  2.13940237e-02 -3.84900905e-02  2.01945025e-02  1.20766172e-02\n",
            " -3.12067149e-03  7.84034468e-03  3.30334972e-03 -4.94357683e-02\n",
            "  5.83886690e-02  3.26137990e-03  4.84483130e-03 -4.50682305e-02\n",
            "  2.45682914e-02  3.55427973e-02 -5.32506183e-02  9.21152830e-02\n",
            "  2.04394981e-02 -3.36951911e-02 -6.19803667e-02 -2.11039372e-02\n",
            "  7.82358795e-02  5.11908121e-02  5.93170337e-02 -1.25138249e-04\n",
            "  4.96350192e-02 -1.55722611e-02 -3.35678132e-03  1.82016175e-02\n",
            " -2.73444112e-02 -1.08772302e-02  1.41476206e-02  1.09877614e-02\n",
            "  4.32554120e-03  8.23311061e-02 -9.85351857e-04  7.58791268e-02\n",
            "  9.44995414e-03  2.37687770e-02  1.61927696e-02  6.24993593e-02\n",
            "  4.75922152e-02 -3.92625527e-03  9.07524452e-02  4.49874513e-02\n",
            " -3.47131379e-02  2.14077290e-02 -3.35604399e-02  4.93849255e-02\n",
            "  1.08669596e-02  2.63447315e-02 -3.26090083e-02  8.00303668e-02\n",
            "  9.29764099e-03  7.16581754e-03 -2.79172305e-02 -3.06821167e-02\n",
            "  4.01061308e-03 -4.93906848e-02 -3.13766720e-03  4.00537960e-02\n",
            " -3.97855081e-02  5.48014119e-02  1.36232920e-05 -8.38373005e-02\n",
            " -1.21547757e-02  3.40949483e-02  3.22403293e-03  6.11845851e-02\n",
            "  5.60067147e-02  9.62868426e-03  2.54616365e-02 -4.64168824e-02\n",
            " -3.98901179e-02  7.68131986e-02  2.28409003e-02 -2.26568524e-02\n",
            " -1.91193614e-02 -6.53029084e-02  4.56781089e-02 -4.43657534e-03\n",
            "  1.49631901e-02 -2.15077940e-02  2.74249981e-03  1.90358274e-02\n",
            "  5.91888502e-02 -2.47569326e-02  3.66144404e-02  5.63083962e-02\n",
            " -8.86448752e-03 -1.74324475e-02 -1.03288016e-03  2.47666314e-02\n",
            "  1.30763054e-02  5.04632853e-02 -5.28496411e-03  5.92397638e-02\n",
            "  6.29906207e-02 -4.36783433e-02 -4.97830249e-02  5.56296855e-02\n",
            " -2.44853850e-02 -8.26754868e-02  2.04910524e-02 -1.06446289e-01\n",
            "  6.64840313e-03  2.97304858e-02 -2.36440394e-02 -8.84616654e-03\n",
            "  2.45556212e-03 -3.35234217e-02  7.52212405e-02 -5.89880198e-02\n",
            " -3.67808491e-02  3.41542847e-02  5.41130826e-02 -1.74904745e-02\n",
            "  1.33920852e-02  4.71682437e-02  1.46116382e-02 -2.12310851e-02\n",
            " -6.55338690e-02  1.23857660e-02  2.76074912e-02 -8.02162010e-03\n",
            " -4.59636636e-02 -8.22441280e-03  9.16955061e-03 -1.56399272e-02\n",
            "  7.54621858e-03  1.58311694e-03 -3.03958263e-02 -5.10671102e-02\n",
            "  1.96313895e-02  1.26263080e-02 -1.51736313e-03  2.02891696e-02\n",
            "  1.37817422e-02  1.49110639e-02  2.50766799e-02 -3.62870470e-02\n",
            "  1.08085470e-02  2.74132611e-03  1.81510597e-02  5.39872162e-02\n",
            " -4.74542119e-02 -4.28730771e-02 -2.89914198e-02  2.13235393e-02\n",
            " -3.85161303e-02  6.31922707e-02 -5.77975921e-02  3.77887371e-03\n",
            " -2.54394077e-02 -1.77221402e-04  9.08246171e-03  1.59095451e-02\n",
            "  4.11799699e-02 -3.94367501e-02 -9.64435469e-03  1.30791226e-02\n",
            "  6.87962249e-02  4.32192720e-02  7.53988861e-04  6.77741393e-02\n",
            "  4.93705869e-02 -3.47822974e-03 -1.06055206e-02  6.72493502e-03\n",
            " -1.39062526e-02  4.88276556e-02 -1.05735194e-02  3.50223086e-03\n",
            "  2.90228007e-03  2.40044240e-02  1.20271696e-02 -2.09797733e-02\n",
            " -2.39111856e-02  3.26579362e-02 -1.01329759e-03 -5.92753664e-03\n",
            " -7.40531320e-03  3.63146723e-03 -2.26698499e-02 -2.21241470e-02\n",
            "  3.86996605e-02  1.72321927e-02  3.85920033e-02 -5.04711568e-02\n",
            " -3.42145115e-02 -4.00443934e-02 -3.57910767e-02 -4.62561324e-02\n",
            "  6.70232177e-02 -4.61649662e-03 -3.29675269e-03  2.08444465e-02\n",
            " -5.14258211e-03 -5.00849895e-02  2.22503915e-02  4.66933772e-02\n",
            "  1.36208758e-02  1.77530069e-02  4.28080093e-03 -2.79332325e-02\n",
            " -1.93421096e-02 -3.87859829e-02 -3.09554469e-02 -6.64134398e-02\n",
            " -1.13433953e-02  1.64267533e-02  1.77629739e-02 -2.28224183e-03\n",
            " -3.30087505e-02 -1.36263284e-03 -2.17934493e-02 -2.67508309e-02\n",
            " -1.26375612e-02  1.61869323e-03 -4.95672598e-02  7.85446092e-02\n",
            "  4.10962887e-02  9.65921115e-03 -1.14643481e-02  1.68858981e-03\n",
            "  5.37662953e-02  2.05532974e-03 -4.11199965e-02  1.46330167e-02\n",
            " -3.75564173e-02 -3.35689783e-02  5.19254804e-03 -6.33088797e-02\n",
            "  3.32963243e-02  8.76118429e-03  1.33859867e-03 -3.95742478e-03\n",
            " -1.61677748e-02  8.26746598e-02  4.75945137e-02 -3.43055725e-02\n",
            "  2.50880960e-02 -3.50977443e-02  3.68657485e-02  4.12650127e-03\n",
            "  4.16017845e-02 -1.35181695e-01 -4.76338193e-02 -1.20025752e-02\n",
            " -3.48891653e-02  3.25455032e-02 -2.93580024e-03 -4.85053752e-03\n",
            " -1.04223713e-01  2.78610419e-02  1.41570149e-02  3.94395217e-02\n",
            " -3.88806835e-02 -1.42463734e-02 -5.19983917e-02  8.92731640e-03\n",
            " -1.99771058e-02 -2.51725148e-02 -3.41299735e-02  1.93041563e-02\n",
            " -5.20207956e-02 -6.72000423e-02 -9.46364645e-03 -1.25585229e-03\n",
            " -5.66048585e-02  2.62098424e-02  9.91584174e-03  4.38286550e-02\n",
            "  2.26643379e-03 -3.11895553e-02 -6.25467747e-02 -3.87793221e-02\n",
            " -6.83939159e-02  4.93721664e-02  5.85507713e-02 -4.08729427e-02\n",
            " -1.98638085e-02 -2.12634560e-02  4.98037562e-02 -4.51748557e-02\n",
            " -2.37141773e-02  2.32674759e-02  1.00594789e-01  9.87118296e-03\n",
            " -1.38015077e-02 -5.21041602e-02  9.08206403e-03  1.72427595e-02\n",
            "  5.91431372e-02  2.62336880e-02 -7.04640616e-03 -1.50032090e-02\n",
            " -3.76658537e-03  6.28258800e-03 -5.23981117e-02 -4.96638566e-02\n",
            "  3.06611080e-02 -3.33646522e-03  2.34911181e-02 -8.58829916e-02\n",
            " -4.62449081e-02  5.59700914e-02  3.09063587e-04  2.01728866e-02\n",
            " -2.98053538e-03  1.76646058e-02  1.54669620e-02 -7.41718039e-02\n",
            "  7.34992698e-03 -1.05015635e-02  2.45247297e-02  1.36879468e-02\n",
            " -1.17803654e-02  4.51544970e-02  3.29039358e-02 -3.50389048e-03\n",
            " -2.71314792e-02 -5.27363792e-02 -4.60164174e-02  2.22848989e-02\n",
            "  2.62271296e-02  5.56151988e-03  1.45789236e-02 -2.97145657e-02\n",
            "  3.57042886e-02  2.22533774e-02  3.89617421e-02 -7.92634487e-02\n",
            " -9.01089516e-03  2.19013132e-02 -5.49054425e-03  8.69956426e-03\n",
            "  4.33031246e-02 -2.12631393e-02  1.13291992e-02 -6.33698925e-02\n",
            "  3.63723636e-02  2.67442726e-02 -6.64251819e-02  1.70499552e-02\n",
            " -2.79691126e-02  2.36355886e-03 -1.81952789e-02  1.52955763e-02\n",
            " -8.50433297e-03  1.16647389e-02 -9.75921750e-02 -2.92093456e-02\n",
            " -5.42547703e-02  3.61235105e-02  3.25117111e-02  8.26975610e-03\n",
            " -2.96499085e-04  1.11556267e-02 -3.85188013e-02  2.36161407e-02\n",
            "  9.85921361e-03  5.73998727e-02  4.86061424e-02 -1.37579199e-02\n",
            " -6.19221339e-03  1.11973006e-02 -3.37175131e-02 -1.10515822e-02\n",
            " -7.08333626e-02 -1.01816598e-02 -3.66010703e-02 -1.55560989e-02\n",
            " -2.13110335e-02 -1.02760708e-02 -4.35733646e-02  5.55186495e-02\n",
            " -3.76548246e-02  5.29252477e-02 -3.45224105e-02 -2.43007811e-03\n",
            "  7.25553632e-02  4.45072725e-03  4.71417420e-02 -9.43884347e-03\n",
            " -1.98979080e-02  5.71900122e-02  8.60541388e-02 -5.25057837e-02\n",
            " -1.39549561e-02  1.17372992e-02  1.33974077e-02 -4.73052524e-02\n",
            " -5.41673936e-02  4.62725945e-02 -2.58969981e-02  1.51415411e-02\n",
            "  3.38944830e-02 -3.78261926e-03 -5.76043315e-02 -1.60082486e-02\n",
            "  2.42738742e-02  3.37360799e-02 -1.96820926e-02 -2.53464375e-02\n",
            " -4.75617126e-02 -5.68755083e-02 -2.28194278e-02  3.83187570e-02\n",
            " -1.78331137e-02  1.35964463e-02  7.85997661e-04  9.74011701e-03\n",
            "  3.34298834e-02 -2.60134004e-02 -7.38565903e-03  3.56451236e-02\n",
            " -2.68532448e-02 -7.53624812e-02 -2.66983807e-02 -4.46457420e-33\n",
            " -3.31646726e-02  1.41703570e-02 -3.92909646e-02 -3.46318446e-02\n",
            " -5.88673959e-03 -1.18212355e-02  1.53951468e-02  1.18473414e-02\n",
            "  1.07757142e-02  3.62140238e-02  7.87954219e-03 -2.31845677e-02\n",
            "  1.07623041e-02  1.72346197e-02  9.54178220e-04  2.83640251e-02\n",
            "  2.37419307e-02 -1.48058375e-02  1.24196196e-03  3.52355768e-03\n",
            "  2.33735796e-02  5.58307655e-02  5.38327880e-02 -3.74079235e-02\n",
            " -2.11805683e-02  1.52712033e-04 -7.27788266e-03  5.50556369e-03\n",
            "  3.05823572e-02  4.54633124e-02 -3.35786715e-02  3.16142254e-02\n",
            " -2.56393314e-03  3.96355428e-02 -1.47573715e-02  5.67167327e-02\n",
            " -5.62787578e-02 -5.04600117e-03  3.56154703e-02 -2.76198816e-02\n",
            " -2.32291892e-02 -4.63291854e-02 -3.70919555e-02 -4.23188917e-02\n",
            "  3.70306931e-02  7.88721535e-03  3.85176092e-02  1.74776383e-03\n",
            "  5.62761398e-03  6.18107338e-03 -6.90268278e-02 -9.42962337e-03\n",
            " -7.74675934e-03  1.68349966e-02  1.22767724e-02  2.26406753e-02\n",
            "  1.21008605e-02  1.11743947e-02  1.21539310e-02 -1.16862329e-02\n",
            " -4.41612378e-02  2.30046958e-02  2.20672451e-02 -5.87504357e-02\n",
            " -3.96428369e-02  6.83135390e-02 -3.29948142e-02 -3.66775058e-02\n",
            " -3.53655033e-02  1.76184867e-02  6.95639802e-03  5.92693239e-02\n",
            "  4.12156470e-02  7.98109397e-02 -5.36562735e-03  1.14238728e-02\n",
            " -2.96388473e-02 -1.15411496e-02  2.22812407e-02  7.93183316e-03\n",
            "  2.60356162e-02  1.28212180e-02  1.71345770e-02 -6.90185558e-03\n",
            " -1.07603427e-02  1.35714635e-02 -9.90795437e-04 -6.16075359e-02\n",
            "  4.40514013e-02 -8.26610194e-04 -2.78340783e-02 -1.23619484e-02\n",
            "  1.34629682e-02 -3.85745242e-02  1.08702201e-03  2.18712538e-02\n",
            " -3.32398415e-02  1.84616055e-02 -5.10114804e-03  3.74664739e-02\n",
            " -3.67544265e-03 -2.19246596e-02 -4.96483222e-03 -9.59838089e-03\n",
            "  2.33591273e-02  1.04876524e-02  4.38722298e-02 -1.51424259e-02\n",
            " -6.30309060e-02  8.23260657e-03 -1.09130749e-02 -4.06409390e-02\n",
            " -6.21691085e-02  2.21326388e-02 -2.71434225e-02  4.05539460e-02\n",
            " -8.09448864e-03 -1.76415453e-03  3.01526655e-02 -5.42263268e-03\n",
            " -4.69822213e-02 -1.73768364e-02  4.11631130e-02  3.20634767e-02\n",
            " -2.22943965e-02 -1.58162061e-02 -4.50720750e-02  5.69486618e-02\n",
            "  4.71595563e-02 -5.78058325e-02  1.32475207e-02 -4.71288944e-03\n",
            "  1.66824464e-07  4.81090546e-02  5.03627956e-02  5.45263179e-02\n",
            "  2.07568631e-02 -1.19080152e-02 -6.37493003e-03  5.26367966e-03\n",
            "  7.21949190e-02 -2.21762881e-02  2.20103282e-02 -9.90471570e-04\n",
            " -1.37163941e-02  6.89207157e-03  2.46913154e-02 -1.39462024e-01\n",
            "  2.56982655e-03 -4.64826375e-02 -4.04967591e-02 -6.08555451e-02\n",
            " -1.53213087e-02  1.36129931e-01  9.45034847e-02  4.25741449e-02\n",
            "  4.67131250e-02 -2.30677761e-02 -1.20965820e-02  3.86673398e-02\n",
            "  2.11645057e-03 -2.51473263e-02 -1.15076350e-02 -3.46506238e-02\n",
            " -2.29533575e-02 -6.33848319e-03 -3.05175744e-02 -1.56236542e-02\n",
            "  1.39514096e-02  3.27477261e-04  2.00321386e-03  4.15107328e-03\n",
            " -2.22924985e-02 -3.62589583e-02 -2.36579478e-02 -1.87817235e-02\n",
            " -1.96289383e-02  4.52125780e-02 -8.12568814e-02 -2.14568917e-02\n",
            " -4.41543311e-02 -2.68475618e-02  2.01974250e-02  2.82990397e-03\n",
            " -1.95011422e-02 -3.45330760e-02  2.26913858e-02  3.78325917e-02\n",
            " -1.02543989e-02 -2.19759904e-03 -8.96744579e-02 -4.50031236e-02\n",
            "  8.09701998e-03 -2.05805562e-02 -2.02998724e-02 -2.09922269e-02\n",
            " -1.79404784e-02  5.81897050e-02 -7.63652008e-03  1.50847323e-02\n",
            "  1.78279755e-34  4.86179404e-02  4.22228537e-02  4.71595228e-02\n",
            "  5.89047521e-02  3.99785116e-02 -5.27071543e-02  1.56905279e-02\n",
            " -5.25070238e-04  1.13651920e-02 -6.56410605e-02 -2.20848955e-02]\n",
            "\n",
            "Sentence: Sentences can be embedded one by one or as a list of strings.\n",
            "Embedding: [ 4.31718193e-02 -5.38701862e-02 -3.78044508e-02  4.27235737e-02\n",
            " -2.35408898e-02  3.44860964e-02  2.89586801e-02  1.92815682e-03\n",
            "  2.41733491e-02 -3.17012779e-02  7.32855946e-02  1.25590302e-02\n",
            "  3.64620574e-02 -2.05251630e-02  2.81973705e-02 -6.87328875e-02\n",
            "  4.22231480e-02  9.31746094e-04  3.54035087e-02  1.41787464e-02\n",
            "  7.83994235e-03  2.31178999e-02 -4.84734494e-03  1.07173799e-02\n",
            "  4.39486839e-03  5.47800958e-03 -3.80338840e-02 -3.05488799e-03\n",
            "  5.72227780e-03 -6.78213686e-02 -4.88007590e-02 -1.45032555e-02\n",
            "  6.68012537e-03 -7.17479512e-02  1.64644837e-06  1.07564405e-02\n",
            " -3.60922813e-02 -2.37057246e-02 -5.22791669e-02  3.46110649e-02\n",
            " -5.42175118e-03  1.62611715e-02  1.96564570e-02  2.25395784e-02\n",
            " -2.25997507e-03  4.06341888e-02  8.17157552e-02  2.48179268e-02\n",
            "  5.31884618e-02  7.82714933e-02 -1.91813745e-02 -1.94087084e-02\n",
            " -2.62805130e-02 -2.44083069e-02  5.49405627e-02  1.90319009e-02\n",
            "  1.60811730e-02 -2.68895365e-02 -8.24695174e-03  7.33444616e-02\n",
            "  1.00123342e-02  2.93315928e-02  3.42861470e-03 -2.13270150e-02\n",
            " -1.62442401e-03 -5.56254806e-03 -7.64879733e-02 -5.85450605e-02\n",
            " -2.82272156e-02  7.51843723e-03  7.11225867e-02  1.95453083e-03\n",
            "  5.45929745e-03  3.22320475e-03  5.12800030e-02 -3.54105718e-02\n",
            " -5.03608398e-02  4.70519029e-02  5.15479455e-03  1.52287185e-02\n",
            " -1.06680188e-02  3.16298492e-02 -9.09038167e-03 -4.01327573e-02\n",
            " -4.35236134e-02 -1.94969568e-02  1.65604819e-02 -4.71168943e-02\n",
            " -3.92091796e-02 -3.07756793e-02 -2.94167586e-02 -4.20826823e-02\n",
            "  2.27072742e-03 -2.78329477e-02  1.69421937e-02  7.74501637e-03\n",
            " -5.23742251e-02 -4.50040139e-02  3.83605249e-02 -4.90786731e-02\n",
            "  5.06619066e-02  1.01615358e-02 -1.25021543e-02 -4.64555155e-03\n",
            " -1.54540055e-02  1.58862602e-02  1.18370363e-02 -3.59232686e-02\n",
            " -7.76225477e-02  3.43358256e-02 -2.14709844e-02 -6.86098486e-02\n",
            " -5.46236411e-02  7.83901513e-02 -3.00702173e-02 -3.37550379e-02\n",
            " -4.04998623e-02  4.80515696e-02  9.53903235e-03  2.31399443e-02\n",
            " -8.16115141e-02 -6.51697442e-03  1.54213104e-02  7.04257712e-02\n",
            " -1.25069106e-02 -2.48267055e-02 -1.71328634e-02  6.13110326e-03\n",
            "  5.44412695e-02 -1.40565978e-02 -6.24514557e-03  3.65788005e-02\n",
            "  7.36230686e-02 -6.05681911e-03 -3.61630619e-02 -1.42203958e-03\n",
            "  4.43166159e-02 -3.14513524e-03  3.18767726e-02 -1.30948164e-02\n",
            " -3.69524509e-02 -4.98028332e-03  1.30016892e-03 -2.05213316e-02\n",
            "  2.06277724e-02  5.93870180e-03 -3.07167275e-03 -3.97513285e-02\n",
            "  4.29890007e-02  6.49802908e-02 -6.76022470e-02  5.41655533e-02\n",
            "  1.52562151e-03 -3.72908860e-02 -4.02427092e-02 -2.28772406e-02\n",
            "  1.31769717e-01  4.87872912e-03  1.39470678e-02  4.92435694e-02\n",
            "  2.49219052e-02 -8.76093004e-03 -5.38768107e-03 -2.65595857e-02\n",
            " -1.19766099e-02 -2.32006814e-02 -2.67433859e-02  5.66913700e-03\n",
            "  2.21722275e-02  4.67294008e-02 -5.78487404e-02  8.22120830e-02\n",
            " -3.36834625e-03  8.09647441e-02  1.41423391e-02  1.02393068e-01\n",
            " -5.76832471e-03 -1.15877781e-02  4.90584895e-02  5.87829612e-02\n",
            "  6.50030747e-02  4.74621579e-02 -2.89463960e-02 -1.76587480e-03\n",
            "  3.32561098e-02  2.91198287e-02  6.03811182e-02  3.73491086e-04\n",
            "  1.06576458e-02 -5.96284755e-02 -7.28601664e-02  2.95079947e-02\n",
            "  9.54464916e-03 -2.71542408e-02 -5.63305095e-02  9.66727501e-04\n",
            " -4.77729291e-02  4.67576869e-02  4.87255212e-03 -6.57519549e-02\n",
            " -1.42248496e-02  3.99872884e-02 -1.09797139e-02  7.68941417e-02\n",
            " -4.00004201e-02  2.96826512e-02  2.81303283e-02 -5.55424877e-02\n",
            "  6.31263386e-03  5.00449911e-02  1.89883914e-02  5.38683832e-02\n",
            " -1.95981693e-02  1.08601190e-02  1.64150428e-02  1.44135421e-02\n",
            "  1.71448812e-02  2.17625257e-02 -4.98862676e-02  1.56104947e-02\n",
            "  4.83737327e-03  1.87053606e-02 -3.18543497e-03  2.66863164e-02\n",
            "  5.55552505e-02 -4.88006026e-02 -3.02929282e-02  2.52110399e-02\n",
            "  1.07265087e-02  1.88270453e-02 -1.50687927e-02  3.43832038e-02\n",
            "  4.15124893e-02  1.37788849e-02 -5.54848835e-02  1.43848425e-02\n",
            " -5.88139519e-02 -6.01676628e-02  2.69856490e-02 -5.46129867e-02\n",
            "  8.14634748e-03 -1.17758401e-02  1.57442857e-02  1.43903506e-03\n",
            " -2.64554638e-02 -4.48876582e-02  4.39732410e-02 -1.06194755e-04\n",
            " -2.25906093e-02  3.00296731e-02  1.97440516e-02  7.44078308e-03\n",
            " -1.93789657e-02  8.09797179e-03  4.34859544e-02 -1.08600296e-04\n",
            " -3.77225578e-02  2.67196149e-02 -4.63157669e-02 -1.53401645e-03\n",
            "  8.05312209e-03 -4.30900976e-02 -2.13848837e-02  1.20185213e-02\n",
            "  8.41399748e-03  2.48273090e-03 -3.09566967e-02 -9.05277282e-02\n",
            " -4.76693623e-02  1.22606354e-02 -1.36467144e-02 -2.63654962e-02\n",
            " -7.65552139e-03  8.72378610e-03  2.65725143e-02  8.40057211e-04\n",
            " -5.55928703e-03 -9.29542072e-03  3.19337286e-02  5.94646633e-02\n",
            "  1.83205772e-02 -7.56546855e-02 -5.59388958e-02 -1.20870778e-02\n",
            " -3.16261090e-02  3.62187251e-02  7.53609231e-03 -6.15653843e-02\n",
            " -2.30459068e-02 -3.51668987e-03  1.23332217e-02 -9.67637636e-03\n",
            "  4.96861488e-02 -8.42256770e-02  1.52395917e-02 -1.82445329e-02\n",
            "  7.70462379e-02  9.28717703e-02  4.03724313e-02  1.11732602e-01\n",
            " -1.03270495e-02 -2.54558027e-02  2.13153809e-02 -1.16186810e-03\n",
            "  2.82604550e-03  5.06967083e-02 -3.13697383e-02 -8.14271905e-03\n",
            "  1.38387512e-02  4.66889553e-02  5.09670787e-02  3.77154052e-02\n",
            " -2.94988807e-02  3.60631905e-02 -2.61165691e-03  2.72258127e-04\n",
            " -6.71806708e-02 -6.54026493e-02 -3.43590789e-02  1.91067513e-02\n",
            "  4.13294099e-02 -1.10970577e-02  4.51951697e-02 -5.93565404e-02\n",
            "  1.06963757e-02 -1.82229895e-02 -5.65814339e-02  1.20386770e-02\n",
            "  4.44773808e-02  1.87050141e-02  1.63810086e-02  5.51150665e-02\n",
            " -2.23332252e-02  2.12861821e-02 -1.20339869e-02  3.26752886e-02\n",
            "  1.47004584e-02 -8.16682260e-03  1.12905195e-02 -3.00620627e-02\n",
            " -2.34345775e-02 -2.68646013e-02 -1.28718873e-03 -7.67189562e-02\n",
            "  2.22609052e-03 -5.89476945e-03  2.63103861e-02  2.07130052e-03\n",
            " -6.91151768e-02 -1.43792182e-02  2.68788096e-02 -3.51539850e-02\n",
            " -2.69611757e-02  2.54717027e-03 -6.48881868e-02  3.18727903e-02\n",
            "  1.70126669e-02 -4.54004444e-02 -1.80616304e-02 -1.61115918e-02\n",
            "  5.70773706e-02 -2.78272573e-03 -6.45585358e-02  7.86597952e-02\n",
            "  2.29075216e-02  6.81841606e-03 -9.11738630e-03 -2.27726288e-02\n",
            " -4.76526655e-02  4.88430671e-02 -2.09892020e-02 -2.43693721e-02\n",
            " -5.01206610e-03  6.70253709e-02  6.91368151e-03  2.25842986e-02\n",
            "  2.51125228e-02 -6.92507625e-03  8.59402865e-03  2.38977186e-02\n",
            "  3.29738334e-02 -1.05310582e-01  1.22094685e-02 -1.22263562e-02\n",
            " -5.73768876e-02  1.84311662e-02  2.97158435e-02 -6.09428994e-02\n",
            " -6.55256063e-02  3.55712995e-02  5.64321782e-03  3.34645738e-03\n",
            " -3.59686762e-02 -8.83417204e-03 -6.97894841e-02  6.89779073e-02\n",
            " -4.88213124e-03  2.23995019e-02 -3.16053554e-02 -7.41183432e-03\n",
            "  3.19351070e-02 -5.18788509e-02  2.11601499e-02 -5.03340401e-02\n",
            "  9.10576154e-03  2.13354323e-02  1.66838020e-02  3.49020213e-02\n",
            " -6.38499707e-02 -6.75625866e-03 -1.27405310e-02 -4.63366471e-02\n",
            " -1.14779770e-02  2.08778344e-02  2.44822130e-02  3.66460346e-03\n",
            " -2.86095333e-03  2.29390450e-02  2.13746261e-02 -3.48900966e-02\n",
            " -3.00387964e-02  4.78870273e-02  5.83370775e-02 -9.70496144e-03\n",
            "  1.38234589e-02 -3.27485688e-02 -8.11501290e-04  9.54227522e-03\n",
            "  1.20401485e-02  1.97230317e-02 -4.74863133e-04 -1.39226038e-02\n",
            " -5.21070361e-02 -1.75592452e-02 -5.41698523e-02 -1.17970034e-02\n",
            " -1.71030592e-02 -3.50195132e-02  3.38661373e-02 -6.76587075e-02\n",
            " -2.27606799e-02  1.95606817e-02  5.50249182e-02  1.22028887e-02\n",
            " -1.75167248e-03  7.22441915e-03  1.16349943e-02 -1.61907859e-02\n",
            " -3.37754898e-02  3.22626941e-02 -2.03813817e-02 -2.33859271e-02\n",
            " -1.29991928e-02 -1.66799258e-02  1.03070885e-02 -1.46030253e-02\n",
            " -7.79070482e-02 -8.25812146e-02 -3.38809863e-02  3.81114595e-02\n",
            "  7.86006637e-03  2.41455045e-02 -2.75714882e-02  1.30866952e-02\n",
            " -7.88593199e-03  1.78651102e-02  5.37323095e-02 -3.01823616e-02\n",
            "  1.69455390e-02  1.19571807e-02  3.52569536e-04  4.90209162e-02\n",
            " -8.57210346e-03  1.71269733e-03  4.83874837e-03 -4.10081223e-02\n",
            " -4.68120985e-02 -2.32559652e-03 -5.16776480e-02  3.10030449e-02\n",
            "  1.60961617e-02 -1.00803478e-02 -3.72483139e-03 -3.53388153e-02\n",
            "  2.95961443e-02  2.89097149e-02 -7.59911686e-02 -5.02980836e-02\n",
            " -2.11783666e-02  3.20462137e-02 -3.84538509e-02  2.45103110e-02\n",
            " -2.04188060e-02  6.02107821e-03 -9.81936697e-03  3.74777764e-02\n",
            "  3.40838991e-02  1.28864441e-02  5.67342751e-02 -8.09703767e-02\n",
            " -8.93608201e-03  1.33352634e-02 -2.51565967e-02  2.58417078e-03\n",
            " -6.51803017e-02  1.34399943e-02 -2.04682462e-02  6.53377734e-03\n",
            "  4.56973491e-03  1.99271273e-02 -6.07340261e-02  1.40691502e-02\n",
            " -5.75334132e-02  9.79801919e-03  3.55393551e-02 -2.45283470e-02\n",
            " -4.73312475e-03 -2.77493019e-02  2.34282743e-02 -8.76351478e-05\n",
            "  7.30441138e-03  1.42028835e-02  4.92807180e-02 -3.16540450e-02\n",
            " -1.34901870e-02  3.08487769e-02  2.80402005e-02 -4.33068871e-02\n",
            " -4.42283973e-02  3.80739197e-02  9.48078305e-05 -4.34896313e-02\n",
            "  1.43868010e-02  2.44328100e-03 -4.84073348e-02  1.08955530e-02\n",
            " -9.87484679e-03  4.59295362e-02  3.96379195e-02 -2.60117278e-02\n",
            "  2.48134509e-02 -5.37149496e-02  5.62824830e-02  8.81359819e-03\n",
            "  5.25076725e-02 -1.47371171e-02 -1.74381137e-02  3.45084183e-02\n",
            "  3.75523455e-02 -4.70167398e-02 -1.94911230e-02  3.82632054e-02\n",
            " -5.67595847e-02 -1.78613677e-03  2.33404394e-02 -5.88216234e-33\n",
            " -4.87188213e-02 -2.76265331e-02 -3.38240489e-02  2.66188327e-02\n",
            " -3.39276679e-02 -8.49191286e-03 -1.91250239e-02  3.00252084e-02\n",
            "  3.40782031e-02  5.11157326e-02 -1.92479938e-02  2.85642501e-02\n",
            "  3.66040282e-02  1.68858543e-02  4.77258191e-02  1.23801949e-02\n",
            "  2.14844178e-02  4.93638741e-04  1.21274088e-02 -5.82144111e-02\n",
            "  1.62954517e-02 -7.14259967e-03  4.80091721e-02  2.51190439e-02\n",
            "  4.60097678e-02 -2.29840446e-02 -2.05696337e-02 -3.22232442e-03\n",
            "  4.00091745e-02  3.52310613e-02 -3.43153998e-02  2.75630318e-03\n",
            " -1.25138909e-02  1.97686236e-02  5.53489337e-03  1.03744589e-01\n",
            "  5.77610172e-03 -5.65426648e-02  4.19558808e-02 -3.78830209e-02\n",
            " -3.93443108e-02 -6.24310002e-02 -2.24398193e-03 -5.46548888e-02\n",
            "  4.56134379e-02 -5.69245312e-03  3.38916928e-02 -1.44447722e-02\n",
            "  2.72107590e-03  1.11192046e-02 -5.00660874e-02 -1.61127280e-02\n",
            "  1.72824331e-03  6.88878298e-02  1.16493050e-02  2.83170994e-02\n",
            "  6.97192084e-03  2.68372390e-02 -7.72086158e-03  2.16828156e-02\n",
            "  1.15182567e-02  8.72832164e-02 -6.27272809e-03 -6.44473583e-02\n",
            " -1.58233587e-02  4.03268225e-02 -1.69728417e-02 -1.61188785e-02\n",
            " -3.75576578e-02  7.02938363e-02 -3.30486484e-02  4.66323793e-02\n",
            "  1.18027823e-02  6.51075244e-02 -1.16979443e-02 -8.28344468e-03\n",
            " -5.46904989e-02 -2.00226773e-02  8.42688954e-04 -8.19525495e-03\n",
            "  2.08356827e-02  1.37454560e-02 -1.29925762e-03 -3.94575298e-02\n",
            " -2.00184621e-02 -1.53721534e-02  1.17271915e-02 -4.40111384e-02\n",
            "  5.39267659e-02 -2.33010873e-02 -2.24211290e-02 -3.65213840e-03\n",
            "  2.92213447e-02  7.56438915e-03 -2.90923100e-02  4.01517898e-02\n",
            " -2.00853311e-02 -1.79863477e-03 -1.26236109e-02  2.51076911e-02\n",
            " -4.69285399e-02 -3.08554638e-02 -3.63343686e-04  6.01787725e-03\n",
            "  3.97509150e-02  1.38547895e-02  2.49774568e-02  1.76976323e-02\n",
            " -9.31572989e-02 -9.83688794e-03  8.44924431e-03 -1.95391141e-02\n",
            " -3.26569416e-02  5.13739092e-03  5.80932433e-03  2.08537206e-02\n",
            " -5.97831607e-03  5.86810336e-02 -1.49496477e-02 -5.72965331e-02\n",
            " -5.98235102e-03  1.95203244e-03  2.72977189e-03  6.06994890e-03\n",
            " -2.00525280e-02 -1.31688165e-02 -4.06229161e-02  5.68997338e-02\n",
            "  4.44968976e-02 -1.24308197e-02  1.96967851e-02  3.80979329e-02\n",
            "  2.30237632e-07  1.10574979e-02  4.79513071e-02  6.18298426e-02\n",
            "  4.40278500e-02  6.17663702e-03  2.58289743e-03  3.38913798e-02\n",
            " -5.32939099e-03 -2.59283409e-02 -1.26143890e-02  2.46495344e-02\n",
            " -1.68772857e-03  1.17901445e-03  2.40443218e-02 -9.77309421e-02\n",
            "  1.97367184e-02 -5.52920513e-02 -6.17424622e-02 -4.87151667e-02\n",
            "  1.11090520e-03  1.18732050e-01  8.13258365e-02  3.32449116e-02\n",
            "  4.38327044e-02 -2.49559321e-02 -3.59626599e-02  1.66319404e-02\n",
            "  5.93765685e-03 -1.43972216e-02  4.46708919e-03 -6.01986311e-02\n",
            " -5.65911531e-02 -8.21538921e-03  5.83059667e-03 -1.69482119e-02\n",
            "  9.58633795e-03  1.46733588e-02  5.05845807e-02  3.06891426e-02\n",
            "  6.60468638e-02 -2.56551728e-02 -2.78858226e-02 -3.19173820e-02\n",
            " -3.39236520e-02  1.49903093e-02 -3.03336028e-02 -6.06493000e-03\n",
            " -4.81772656e-03  1.72137003e-02 -8.23370460e-03  1.55547718e-02\n",
            "  2.69106701e-02  5.44309942e-03 -1.06899524e-02 -7.82140996e-03\n",
            " -4.44506928e-02  2.55874135e-02 -5.74760661e-02 -2.05443110e-02\n",
            " -3.07849888e-02 -1.57854930e-02 -7.07536051e-03 -4.21312414e-02\n",
            "  3.79934236e-02  6.27765357e-02 -7.67793274e-03 -3.18353474e-02\n",
            "  1.99277746e-34  1.04834093e-02 -3.39326560e-02  3.93821299e-02\n",
            "  5.53065650e-02  9.42169689e-03  1.09728258e-02 -4.91939336e-02\n",
            "  2.95024160e-02 -8.85374099e-03 -5.96248657e-02 -2.37825699e-02]\n",
            "\n",
            "Sentence: Embeddings are one of the most powerful concepts in machine learning!\n",
            "Embedding: [-2.98611615e-02 -1.37522090e-02 -4.75402176e-02  2.72126738e-02\n",
            "  3.40053886e-02  3.16466093e-02  4.26964164e-02  3.29799554e-03\n",
            "  4.35718149e-02  2.53836531e-02  3.02528758e-02  3.21130976e-02\n",
            " -3.99912931e-02  1.28761912e-02  6.70219660e-02 -7.92899951e-02\n",
            "  4.68772240e-02  2.40267050e-02 -2.07997710e-02 -1.07433284e-02\n",
            " -1.19410604e-02 -5.39291315e-02  4.21055295e-02  2.23588385e-02\n",
            " -2.98949685e-02  8.35979078e-03  1.58385150e-02 -4.80236448e-02\n",
            "  1.88440620e-03 -1.67521387e-02 -2.15628538e-02 -3.88488360e-02\n",
            "  3.06274295e-02  4.20525968e-02  1.69483371e-06 -1.86929014e-02\n",
            " -1.24558546e-02  1.32128904e-02 -4.89039570e-02  1.34745976e-02\n",
            "  2.28873324e-02  8.81785806e-03  8.64930078e-03 -2.00950000e-02\n",
            " -3.15217748e-02 -2.53433287e-02  7.57318512e-02  3.62447016e-02\n",
            "  1.25290258e-02  3.09695136e-02  4.50759567e-03 -3.50042433e-02\n",
            " -4.42457036e-04 -9.76647157e-03  6.04545549e-02  4.03472222e-02\n",
            "  1.10734832e-02  6.56198896e-03 -5.84597327e-03  3.79785383e-03\n",
            " -4.46914807e-02  1.76405106e-02  2.45916825e-02 -3.60042253e-03\n",
            "  1.02473363e-01  3.73758785e-02  6.13311958e-03 -2.24676803e-02\n",
            "  1.46482522e-02  5.00536561e-02 -2.29907259e-02  1.12924809e-02\n",
            " -3.10552251e-02 -1.49508938e-02 -2.53131590e-03  3.20945159e-02\n",
            " -4.67056148e-02 -4.85886745e-02  2.98306178e-02  6.44215718e-02\n",
            " -3.12613174e-02  3.57407257e-02  4.16526757e-02 -5.52517548e-02\n",
            " -8.74624681e-03 -2.18630042e-02 -1.12744272e-02 -2.14436240e-02\n",
            " -1.32823642e-02 -2.04865988e-02 -1.00576291e-02  3.54763865e-02\n",
            " -7.47610582e-03 -3.70188318e-02  5.77893220e-02 -2.18169112e-02\n",
            "  4.36223811e-03  2.04380434e-02  3.36814784e-02 -4.92800996e-02\n",
            "  4.82793339e-02 -1.81002531e-03 -1.05118677e-02  4.13323231e-02\n",
            " -6.79834038e-02  1.75715946e-02 -4.43412475e-02  9.90841165e-03\n",
            " -3.81809063e-02  1.10827768e-02 -5.07280082e-02 -2.17451118e-02\n",
            " -1.03836292e-02  4.60332856e-02  1.55862542e-02 -4.21366692e-02\n",
            " -2.72146296e-02  3.22818942e-02 -4.24739793e-02  2.71207206e-02\n",
            " -7.41061717e-02  4.20107581e-02  2.02438347e-02  7.31811300e-02\n",
            " -8.97695404e-03 -2.31160428e-02 -3.93559337e-02 -1.46008758e-02\n",
            " -3.30910310e-02  1.12239504e-02  2.58571608e-03 -4.36851615e-03\n",
            "  1.85855366e-02  2.69934721e-02 -1.67215262e-02  3.69569734e-02\n",
            "  4.44489196e-02 -2.21723653e-02  6.72961446e-03  1.22935493e-02\n",
            "  1.71758719e-02 -2.36478099e-03  3.72263342e-02 -2.22870428e-02\n",
            "  2.94603277e-02 -2.33691521e-02  5.38469292e-03 -3.06581948e-02\n",
            " -2.38920078e-02 -2.63614226e-02 -2.01788712e-02  1.11245610e-01\n",
            " -1.99836083e-02 -3.54030766e-02  3.84143740e-02  2.53069550e-02\n",
            "  1.99551564e-02  5.53518161e-02 -1.99332908e-02 -2.16719159e-03\n",
            "  4.91092876e-02 -4.03530635e-02 -1.16977133e-02 -5.33113144e-02\n",
            "  8.29587784e-03 -5.08251674e-02 -2.65503973e-02 -1.53242880e-02\n",
            "  5.78816188e-03  2.46578455e-03 -3.44449654e-02 -1.85128290e-03\n",
            " -3.95730548e-02 -2.71690749e-02  4.93568443e-02  8.38368908e-02\n",
            "  5.43492176e-02  8.22261423e-02  1.23894550e-02 -4.79798531e-03\n",
            "  7.77383510e-04  2.98486706e-02 -1.85583923e-02  5.98795451e-02\n",
            " -6.82795746e-03  9.78120137e-04  2.85485629e-02 -7.64621003e-03\n",
            " -1.86620243e-02 -2.69287620e-02 -2.90332660e-02 -1.37871904e-02\n",
            " -2.57597747e-03 -2.20173635e-02 -1.70821529e-02 -3.81842740e-02\n",
            "  2.21505575e-02 -3.59234065e-02 -1.19439848e-02 -3.18307616e-02\n",
            " -4.80801053e-02  9.77636036e-03 -1.04860205e-03  4.15370874e-02\n",
            " -1.10974386e-02 -4.72830534e-02  1.90984756e-02 -5.31177334e-02\n",
            "  2.11325325e-02 -2.53069284e-03  5.61054908e-02 -1.33795785e-02\n",
            " -5.95863350e-03 -1.20308287e-02  4.63929772e-02 -2.81908978e-02\n",
            "  2.25355364e-02 -2.50471290e-03 -3.52452807e-02  2.55495682e-02\n",
            "  9.10396688e-03  3.25214816e-03  2.55972496e-03 -1.25624510e-02\n",
            " -3.51496525e-02 -4.28946577e-02 -2.32333806e-03  2.41021216e-02\n",
            " -5.16839186e-03  1.68740358e-02  5.52645605e-03  2.36792006e-02\n",
            "  5.65164499e-02 -3.47869098e-02 -6.34518042e-02 -7.45626027e-03\n",
            " -1.78448465e-02  5.35897501e-02  2.67291758e-02 -8.74199644e-02\n",
            "  1.04196360e-02 -4.13941889e-04 -3.04447510e-03  9.14244819e-03\n",
            "  2.91530192e-02 -5.81831373e-02  6.83463141e-02 -4.08618338e-02\n",
            " -9.09787323e-03 -3.40770334e-02  3.52411456e-02 -1.02628255e-02\n",
            " -5.72466524e-04 -2.73451488e-03  1.59635805e-02  4.49064514e-03\n",
            " -2.09051650e-02  3.02770045e-02  2.46119164e-02 -1.44068217e-02\n",
            "  1.73269827e-02  1.99038861e-03  4.23051901e-02 -2.39176489e-02\n",
            " -3.25547233e-02 -1.45940054e-02  3.95101048e-02 -6.04650192e-02\n",
            " -3.02065630e-02  1.67189669e-02 -2.26816870e-02 -2.61954647e-02\n",
            " -5.51319532e-02  1.44908438e-02 -1.99246276e-02  3.99747258e-03\n",
            "  3.12610157e-02 -4.90727648e-02 -9.49762587e-04  5.39497249e-02\n",
            " -9.10082553e-03 -2.69485917e-02 -3.63158733e-02 -1.38436779e-02\n",
            " -4.45622243e-02  5.49359098e-02  2.17597815e-03  2.23441934e-03\n",
            " -5.23013668e-03 -1.47894118e-02  3.60592194e-02  1.45263411e-02\n",
            "  8.39192793e-03 -6.10360578e-02 -7.89953209e-03 -2.98294774e-03\n",
            "  3.56567348e-03  8.33992064e-02 -2.61215270e-02  8.06722119e-02\n",
            "  3.63051519e-03  1.69974845e-02  2.58604214e-02  1.09442824e-03\n",
            " -4.57063317e-02  5.55678681e-02  2.00643297e-02  4.76660840e-02\n",
            " -4.91053089e-02 -1.86081417e-02  3.34404930e-02 -2.57310402e-02\n",
            " -3.16365249e-03  7.21444264e-02 -1.61519460e-02 -1.33933090e-02\n",
            " -6.06294237e-02 -2.82187071e-02 -8.91928840e-03 -2.71171238e-03\n",
            "  8.04919284e-03 -4.95209545e-02  7.89434537e-02  2.76427977e-02\n",
            " -5.42582897e-03 -3.06726829e-03 -4.11827080e-02  1.39171779e-02\n",
            "  3.04253157e-02  1.02856401e-02  1.06679136e-02 -5.56554385e-02\n",
            " -1.75083298e-02  2.03868803e-02  8.43314081e-03  3.82470898e-02\n",
            " -3.89099903e-02 -1.61303412e-02  3.18059921e-02 -7.32969940e-02\n",
            " -1.76502373e-02 -4.79874052e-02 -5.55042140e-02 -5.00750076e-03\n",
            "  4.46798978e-04  3.57332937e-02 -8.24674149e-04 -3.34324315e-02\n",
            " -3.32416743e-02 -2.46460270e-02  2.15332285e-02  3.90860531e-03\n",
            "  2.53471453e-02  6.02994114e-03 -7.81607628e-03  1.23765925e-02\n",
            " -1.71039272e-02  2.68102791e-02  2.83670239e-03 -1.27643198e-02\n",
            "  1.00510873e-01  1.03580691e-02 -3.55142914e-02  1.56616364e-02\n",
            " -9.85949561e-02  4.58441749e-02 -3.15230899e-02 -2.35781278e-02\n",
            " -2.78350711e-02 -7.75875378e-05 -2.82363389e-02 -1.92918591e-02\n",
            "  1.87389944e-02  5.71941249e-02  2.56912094e-02 -3.20030376e-02\n",
            "  1.99074261e-02 -3.15819904e-02 -4.02062275e-02  5.77630661e-02\n",
            "  1.72974002e-02 -5.37014343e-02 -1.25325732e-02 -1.45484423e-02\n",
            " -5.76175004e-02  1.09727690e-02 -2.04728488e-02  2.85541043e-02\n",
            " -5.04399762e-02  4.36991155e-02  1.75711047e-02 -1.02344044e-02\n",
            " -9.69772041e-02 -2.99995132e-02 -2.86679287e-02  2.24936623e-02\n",
            " -1.68121178e-02 -1.43673755e-02 -8.79590865e-03 -1.69043746e-02\n",
            "  2.41558235e-02 -6.53192624e-02 -4.10799943e-02 -2.34058686e-02\n",
            " -6.76065087e-02 -1.55690713e-02  3.62358317e-02  7.83160031e-02\n",
            " -4.97517176e-02 -7.08547980e-02 -5.01179025e-02 -8.56435159e-04\n",
            "  5.44897327e-03  4.34699422e-03  9.88052860e-02 -2.16415636e-02\n",
            " -1.87750850e-02  1.15070017e-02  2.63996273e-02  1.65235698e-02\n",
            " -2.24057846e-02 -4.31826822e-02  1.31803945e-01 -2.97034755e-02\n",
            "  2.65934225e-02 -1.38888676e-02 -1.67003851e-02  3.44146229e-02\n",
            " -8.94356798e-03  6.16001263e-02 -3.42303254e-02  2.46419152e-03\n",
            " -8.14090855e-03  5.80325760e-02  5.24208322e-02 -1.53281540e-02\n",
            "  4.01382819e-02  1.51407821e-02 -3.01468698e-03 -4.97021489e-02\n",
            " -4.24302649e-03  5.77288419e-02  3.17873694e-02  4.74007875e-02\n",
            "  2.95218136e-02 -1.50122950e-02 -2.47945730e-02 -7.11502284e-02\n",
            "  2.06847545e-02  3.11487522e-02 -5.86067792e-03  1.62786935e-02\n",
            " -3.93682383e-02  5.46505600e-02  3.26595046e-02 -1.87021531e-02\n",
            " -9.79863331e-02  4.33768472e-03 -5.58157302e-02 -1.34620974e-02\n",
            "  2.88453996e-02  1.58748161e-02 -3.32564898e-02  1.44415814e-03\n",
            " -5.51108010e-02  8.24674368e-02  2.38846000e-02 -2.04837732e-02\n",
            " -4.78586787e-03  3.78722884e-02 -4.87562940e-02  3.44646871e-02\n",
            "  1.10358084e-02  1.12450747e-02  1.33262491e-02 -3.46375667e-02\n",
            " -6.92220926e-02  7.30126025e-03 -6.57012314e-03  1.73203330e-02\n",
            "  5.23057114e-03  4.48132232e-02  3.89853194e-02 -1.99274216e-02\n",
            " -1.80920698e-02  3.25937532e-02 -2.02026237e-02  4.86206700e-04\n",
            " -8.88755545e-03 -1.91348363e-02  2.50685718e-02  4.74020317e-02\n",
            "  2.18658079e-03 -1.69988275e-02  3.62671576e-02  3.46244965e-03\n",
            "  4.21938486e-03  8.04172531e-02  3.10627613e-02 -1.04941835e-03\n",
            " -3.55466641e-02  4.34836932e-02 -3.06218714e-02 -3.03192288e-02\n",
            " -4.13175076e-02 -1.05257835e-02 -2.35242844e-02 -1.86771844e-02\n",
            "  4.42935573e-03  5.45056127e-02 -6.05017617e-02  2.48421840e-02\n",
            " -3.36966217e-02 -4.54169065e-02 -2.63173338e-02  6.98052673e-03\n",
            "  6.92871660e-02 -2.04491280e-02 -1.96813494e-02 -9.72562283e-03\n",
            " -1.21564772e-02  7.89341982e-03  1.84745819e-03 -6.93650916e-02\n",
            "  2.43357103e-02  4.00609225e-02  3.44013125e-02 -2.84281913e-02\n",
            " -1.09431241e-02  1.38742812e-02 -4.40585287e-03  1.19345612e-03\n",
            " -8.81164819e-02  1.15930764e-02 -2.56351605e-02  5.57525903e-02\n",
            "  1.26946136e-01  5.39566018e-02 -1.41436392e-02  1.27196210e-02\n",
            " -1.32236034e-02 -5.94484396e-02  2.86704693e-02  2.57284883e-02\n",
            " -8.33775196e-03  8.17429216e-04  5.93050849e-03  3.29109989e-02\n",
            "  4.12751958e-02 -5.77958021e-03 -1.71124320e-02  1.06227808e-02\n",
            " -2.19601393e-02 -4.97207195e-02  2.53765434e-02 -5.60257579e-33\n",
            " -1.35397352e-02 -3.77959087e-02 -2.67923623e-03 -3.69689486e-04\n",
            " -1.98267996e-02  5.47052221e-03  6.02684729e-03  1.93068981e-02\n",
            "  3.87984910e-03  2.97698658e-02 -1.88228786e-02  2.20034621e-03\n",
            "  8.17020424e-03  1.61965508e-02  3.17528620e-02 -6.83420664e-03\n",
            "  2.19252091e-02  4.37734532e-04  2.96859164e-02 -2.62557324e-02\n",
            "  6.49381056e-03  3.56025174e-02  1.58609450e-03 -4.76583838e-02\n",
            " -5.26238643e-02  3.78274024e-02  3.54341641e-02 -3.10347937e-02\n",
            "  7.96400756e-03  5.48469201e-02 -3.56443301e-02  9.10133403e-03\n",
            " -9.45984386e-03 -4.63287197e-02 -1.63906608e-02  6.32453188e-02\n",
            " -1.38588622e-02 -5.95724732e-02 -1.57991052e-02  2.01887432e-02\n",
            " -1.98293459e-02 -3.49211507e-02  2.27938127e-02 -5.91621660e-02\n",
            "  4.18854430e-02  1.20744819e-03  5.19158654e-02 -1.88435949e-02\n",
            " -3.12101711e-02  2.34932993e-02 -7.41029829e-02 -2.76645325e-04\n",
            " -1.51720373e-02  6.11714087e-02  1.25065133e-01 -1.28459297e-02\n",
            " -1.12671442e-02  1.51749223e-03 -8.09153542e-02  1.12689175e-02\n",
            " -1.97574161e-02  2.74268202e-02  9.40990541e-03 -9.58757848e-03\n",
            "  2.54850890e-02  6.81658909e-02 -1.83452927e-02 -1.00963950e-01\n",
            " -9.45247803e-03 -5.27006062e-03  1.98684093e-02  9.80847478e-02\n",
            "  3.15633118e-02  5.30422926e-02  3.75123024e-02 -6.64209053e-02\n",
            " -5.92880435e-02 -1.57073718e-02  1.76609475e-02 -5.81073873e-02\n",
            "  2.23230850e-02  1.29869804e-02 -3.30261216e-02  9.96865099e-04\n",
            " -9.87089425e-03 -3.12955193e-02  2.28525070e-03 -4.91250232e-02\n",
            "  1.47693492e-02 -1.83367580e-02 -4.16305251e-02  3.76897268e-02\n",
            "  3.35410200e-02 -7.97111541e-02  4.01300378e-02  1.59071442e-02\n",
            "  5.06089302e-03  4.28808592e-02  2.29760259e-02 -4.13350984e-02\n",
            " -3.10503226e-02 -5.26404902e-02 -4.95405085e-02 -2.94256508e-02\n",
            "  5.94924279e-02 -2.59803087e-02  3.02497428e-02  8.80410802e-03\n",
            " -4.84467186e-02 -2.00851355e-02  9.82172880e-03 -7.89193735e-02\n",
            "  4.52884007e-03 -9.34895221e-03  9.23312176e-03 -3.17362249e-02\n",
            "  2.10833047e-02  6.37123501e-03  3.36347744e-02  3.83613594e-02\n",
            " -4.55528051e-02  1.08117843e-03 -9.83321294e-03  7.70189054e-03\n",
            " -2.87617501e-02 -1.74959432e-02 -4.27811267e-03  2.81287301e-02\n",
            "  4.97339182e-02 -7.45570809e-02 -1.07008554e-02 -7.66055938e-03\n",
            "  2.33968819e-07  1.52483070e-02  8.39613974e-02  3.67242955e-02\n",
            " -3.69249247e-02  3.64752337e-02  4.26422060e-02 -4.39824676e-03\n",
            "  1.78133752e-02 -2.67076921e-02 -7.13898800e-03  5.59975952e-02\n",
            "  3.13965939e-02  2.13439390e-03  3.90370674e-02 -8.78528729e-02\n",
            " -2.21659914e-02 -2.47735735e-02 -1.18190618e-02 -7.89707340e-03\n",
            " -2.08857376e-02  4.30556312e-02  1.07643515e-01  4.40618806e-02\n",
            "  1.47963315e-02  2.44862903e-02 -3.86268087e-02  1.80743653e-02\n",
            " -1.47838157e-03  7.74167255e-02 -4.19565663e-02 -3.80529054e-02\n",
            "  3.61258015e-02  1.59038266e-03  1.95323452e-02 -2.00080238e-02\n",
            "  4.22538482e-02  3.06111630e-02 -3.53699736e-03  5.93108824e-03\n",
            " -2.23223511e-02 -2.07131281e-02 -3.62916989e-03  1.74653921e-02\n",
            " -4.08758484e-02  5.91594912e-02 -5.89099042e-02 -3.96754332e-02\n",
            " -3.33528817e-02  1.02162044e-02  6.97289826e-03  7.70389810e-02\n",
            " -1.86911635e-02 -1.82568394e-02 -2.42318958e-02 -3.40701244e-03\n",
            " -3.60554941e-02  4.33389544e-02 -3.48602720e-02  5.27768284e-02\n",
            "  2.89709978e-02 -4.98462878e-02 -1.94749124e-02  1.16398185e-02\n",
            " -3.04614920e-02  8.04638043e-02  6.56248182e-02 -2.84533035e-02\n",
            "  1.81615278e-34 -4.19155834e-03 -2.57882401e-02  5.17319962e-02\n",
            "  4.94420715e-02  1.32476697e-02 -4.21994030e-02 -1.12458607e-02\n",
            " -2.61520464e-02  5.51130548e-02  2.20024977e-02 -2.51170527e-02]\n",
            "\n",
            "Sentence: Learn to use embeddings well and you'll be well on your way to being an AI engineer.\n",
            "Embedding: [-2.20730901e-02  2.08950527e-02 -6.03005923e-02  8.43949243e-03\n",
            "  4.37650681e-02  1.55070797e-02  4.99907285e-02 -3.03232390e-02\n",
            "  4.94784415e-02  2.35511623e-02  3.29350606e-02  1.53877335e-02\n",
            " -6.68355003e-02  1.11002862e-01  6.92676604e-02 -2.31889058e-02\n",
            "  3.79103348e-02 -4.94145928e-03 -1.57800876e-02 -3.45476381e-02\n",
            " -2.65053026e-02 -2.47879587e-02 -1.86141264e-02  3.00361458e-02\n",
            " -2.81186569e-02 -8.75141565e-03 -3.30770877e-03 -2.06116512e-02\n",
            "  1.03315664e-02 -1.51483212e-02 -3.48331220e-02 -2.63248254e-02\n",
            "  2.06907690e-02  3.79109345e-02  1.81912878e-06 -2.44291592e-03\n",
            " -1.80561387e-03  5.61762601e-03 -2.79870126e-02  1.54703129e-02\n",
            "  3.06457151e-02  3.72601748e-02 -1.55611252e-02  2.54414286e-02\n",
            " -6.42072409e-02  3.16352919e-02  6.63442463e-02  3.80970351e-02\n",
            "  5.57844974e-02  5.31660393e-02 -9.69294738e-03 -3.61424387e-02\n",
            "  3.72435153e-02 -4.67828149e-03  5.14575914e-02  1.00057628e-02\n",
            "  4.90292255e-03  1.41562922e-02  4.95099761e-02  3.32949497e-03\n",
            " -3.21100689e-02  4.42387573e-02  3.27415615e-02 -7.90621527e-03\n",
            "  1.07809797e-01  7.32944459e-02  3.36702466e-02 -4.28345725e-02\n",
            "  1.05966330e-02  2.05653887e-02 -2.02670153e-02  1.04964459e-02\n",
            " -1.97615456e-02 -2.89213040e-05 -2.61862110e-02 -1.85174067e-02\n",
            " -3.44269313e-02 -4.08620834e-02  2.32571792e-02  2.14195978e-02\n",
            "  1.31321438e-02 -3.27211171e-02 -1.91425867e-02 -2.86572147e-02\n",
            " -1.16859190e-02  1.21910647e-02  1.05248699e-02 -3.39584760e-02\n",
            "  3.08911037e-03 -4.44888733e-02  2.65104696e-02  1.09536657e-02\n",
            "  2.51450706e-02 -6.48833346e-03  4.54362947e-03 -2.02784855e-02\n",
            " -1.03216078e-02  2.06589848e-02 -1.65313873e-02 -2.45612226e-02\n",
            "  5.47549836e-02  2.68261526e-02  2.95510143e-02  3.86755019e-02\n",
            " -7.76720643e-02  3.80055793e-02 -2.98364293e-02  7.96886086e-02\n",
            " -3.00942920e-02  7.57843116e-03 -6.89826757e-02 -2.92667132e-02\n",
            " -2.35580634e-02  3.48198861e-02  2.52938755e-02 -4.53817174e-02\n",
            " -1.57939326e-02  4.39031161e-02 -4.04335931e-02  8.32530390e-03\n",
            " -2.84664948e-02  4.94933650e-02  2.41277311e-02  3.02192494e-02\n",
            " -4.99590337e-02 -5.94532788e-02 -3.70175913e-02  1.30330212e-02\n",
            " -3.36469226e-02  3.45589891e-02 -1.44523010e-02  2.57640444e-02\n",
            "  4.61187446e-03  2.21551750e-02 -4.93460521e-03  9.66004655e-02\n",
            " -2.72446265e-03  5.65331720e-04 -3.24248523e-02  1.31681720e-02\n",
            "  4.41607200e-02 -7.03055691e-03  6.84260949e-02 -2.28166748e-02\n",
            " -2.81032547e-03 -4.23883088e-02 -1.33632170e-02 -5.96738532e-02\n",
            " -6.96127769e-03 -2.31900979e-02 -3.78851108e-02  9.80186835e-02\n",
            " -2.21728701e-02 -2.30062716e-02  3.22815068e-02  8.21806025e-03\n",
            " -7.04117585e-03  4.84079644e-02  4.23291586e-02 -2.59714690e-03\n",
            "  1.20535435e-04  1.67414378e-02  2.91197747e-02 -1.28740864e-02\n",
            " -2.41077859e-02 -3.29320431e-02 -3.50295799e-03 -3.19321752e-02\n",
            " -2.64171846e-02  2.30404232e-02  1.11637255e-02 -9.95993987e-03\n",
            " -1.75901148e-02 -2.00276799e-03  1.21594863e-02  4.67823744e-02\n",
            "  5.20882271e-02  7.21172020e-02  1.94978192e-02  1.15072355e-02\n",
            "  5.94161265e-03 -1.47834057e-02 -2.87724044e-02  6.72666505e-02\n",
            " -1.68758556e-02  5.25875809e-03 -3.39739099e-02  5.24596423e-02\n",
            " -2.59792972e-02 -4.41379957e-02  1.47448620e-03 -1.06599070e-02\n",
            " -1.51859568e-02 -1.55879592e-03  1.81504488e-02 -4.85410951e-02\n",
            "  3.67902080e-03 -6.59313351e-02 -1.49418423e-02 -3.23529951e-02\n",
            " -2.79948972e-02  1.71856657e-02 -7.87080824e-03  4.65692207e-02\n",
            "  1.47123523e-02 -7.40438998e-02 -6.52104169e-02 -5.22735044e-02\n",
            " -1.82343591e-02  5.20859286e-02  3.06304451e-02 -2.36037709e-02\n",
            "  2.42384635e-02 -1.83940250e-02 -4.83010476e-03 -2.13386118e-02\n",
            "  1.56584028e-02  9.87342931e-03 -4.25561741e-02  6.00459753e-03\n",
            " -3.14499275e-03  4.51517990e-03 -1.52776390e-03  1.13731371e-02\n",
            " -6.96354359e-02 -3.37258205e-02  1.33406883e-02  4.87286784e-03\n",
            " -7.81485159e-03  4.78049442e-02 -1.59711801e-02  3.14606503e-02\n",
            "  5.15920557e-02 -4.05121893e-02 -5.06460853e-02  9.99934133e-03\n",
            " -2.00729929e-02  4.21552472e-02  3.03183068e-02 -1.00431226e-01\n",
            " -4.12020087e-02  3.43990661e-02  3.29209492e-02  1.07408001e-03\n",
            "  3.70962434e-02 -6.94237277e-02  6.52393475e-02  8.31661746e-03\n",
            "  1.68036148e-02 -2.60706078e-02  8.14502966e-03 -1.48009202e-02\n",
            " -2.65670735e-02  3.29321809e-02 -7.37515232e-03  7.23975152e-03\n",
            " -2.69330498e-02  1.71755292e-02 -2.28083134e-02 -4.75341966e-03\n",
            "  2.88568903e-02  1.30799512e-04  5.44129014e-02 -1.43378861e-02\n",
            "  1.89891830e-02 -1.32732205e-02  4.01178151e-02 -7.29275718e-02\n",
            " -2.41210721e-02  3.16217318e-02 -1.68014150e-02  8.47541355e-03\n",
            " -5.23940846e-02 -1.43882986e-02 -1.46156447e-02  6.39907736e-03\n",
            "  2.15113666e-02 -5.18960394e-02 -4.30576205e-02  2.34075822e-02\n",
            "  2.30274582e-03 -2.48434171e-02 -4.38243523e-02 -2.16570273e-02\n",
            " -6.91595599e-02  1.76770259e-02  2.12066602e-02 -2.20294166e-02\n",
            " -1.06772808e-02  9.57421772e-03  2.21988745e-02  5.51471263e-02\n",
            "  1.03682997e-02 -8.14824402e-02 -7.94707518e-03 -1.85866784e-02\n",
            "  1.20494505e-02  7.51403049e-02 -1.41215734e-02  8.83924440e-02\n",
            "  3.12628560e-02  8.12257826e-03 -2.29444131e-02  3.96010913e-02\n",
            " -2.00168118e-02  9.16027576e-02 -2.06839070e-02  5.84991015e-02\n",
            " -4.32368740e-02 -4.74742660e-03 -9.51887015e-03  5.42948162e-03\n",
            "  1.19159944e-04  6.15376197e-02 -1.68788224e-03 -4.66320664e-02\n",
            " -2.01405529e-02  1.32405749e-02  9.70683340e-03  2.73847859e-02\n",
            "  3.06639690e-02  6.71582948e-03  8.71220753e-02 -1.97372329e-03\n",
            "  8.18296988e-03  5.44357719e-03 -5.76206036e-02  1.34820463e-02\n",
            "  5.06295450e-03 -2.10569035e-02  1.25937089e-02 -5.49785560e-03\n",
            " -1.44645432e-02 -2.92567294e-02  5.53298444e-02 -2.60099135e-02\n",
            " -2.82453815e-03 -2.30901707e-02  8.89709685e-03 -2.61565428e-02\n",
            "  9.08626942e-04 -6.16203733e-02 -7.56419748e-02 -1.05931852e-02\n",
            " -1.19563704e-02  6.71458393e-02 -1.96234770e-02 -5.00934087e-02\n",
            " -3.91229019e-02 -3.07007860e-02  7.18906596e-02  9.29244049e-03\n",
            " -6.34385878e-03  7.86979450e-04 -1.36484047e-02  2.87188832e-02\n",
            "  4.01225016e-02  1.28036849e-02  1.77381244e-02 -4.75977408e-03\n",
            "  5.47173433e-02  1.10807235e-03 -2.25790516e-02 -2.80286395e-03\n",
            " -1.13696039e-01  2.55904533e-02  4.00441437e-04 -4.39811535e-02\n",
            "  1.36319147e-02 -1.54137444e-02 -4.99015450e-02 -2.32889690e-02\n",
            " -1.62987679e-03  3.95835452e-02  1.89040601e-02 -3.02703399e-02\n",
            "  2.71438044e-02  1.05672539e-03 -4.21069525e-02  3.71961221e-02\n",
            "  3.54258642e-02 -6.98274821e-02 -2.20937524e-02 -4.01496328e-02\n",
            " -1.90164503e-02 -2.69835237e-02 -1.51213314e-02  3.33365425e-02\n",
            " -9.74889845e-02  1.73102375e-02  6.21019304e-03 -2.59427750e-03\n",
            " -1.10152945e-01 -6.10186532e-02 -1.36549668e-02 -1.35035226e-02\n",
            " -6.72573224e-02 -4.05122340e-03 -6.64987648e-03  3.86565458e-03\n",
            "  9.43470560e-03 -3.86636704e-02 -1.93592608e-02  1.34933833e-02\n",
            " -4.58106920e-02  6.06737025e-02  6.06380068e-02  4.85596843e-02\n",
            " -4.56088334e-02 -5.71936667e-02 -1.55094964e-02  3.40963006e-02\n",
            "  9.48158791e-04 -9.94346291e-03  2.84657702e-02 -3.29024270e-02\n",
            " -2.83211172e-02  3.19907740e-02  2.61298660e-02 -2.74054818e-02\n",
            " -1.36353364e-02  7.47712702e-03  1.19430318e-01 -4.45806906e-02\n",
            "  1.07670883e-02 -8.69424269e-02 -2.19550747e-02  1.83874853e-02\n",
            " -1.06521258e-02 -1.89242978e-02 -3.06513794e-02 -3.04701887e-02\n",
            " -3.22214440e-02  4.12151590e-02  8.95752478e-03 -2.73180194e-02\n",
            "  9.39994399e-03 -9.57133772e-04 -1.94010008e-02 -4.92623486e-02\n",
            " -9.18890163e-03  4.66894470e-02  5.41892834e-02  2.21609380e-02\n",
            " -2.86351740e-02  5.20295575e-02  2.47589014e-02 -7.14268684e-02\n",
            " -1.26206502e-02  7.35517964e-03  2.13784296e-02  2.93514319e-02\n",
            " -2.57651284e-02  5.20562679e-02 -2.74891946e-02 -3.10242735e-02\n",
            " -9.02879909e-02  6.10371493e-02 -5.22609726e-02  2.13111807e-02\n",
            "  4.41735014e-02  3.23371030e-02  1.75648127e-02 -2.39519924e-02\n",
            " -2.69709285e-02  5.11278287e-02  2.69065686e-02 -4.51932363e-02\n",
            "  2.52657826e-03  2.44934168e-02 -2.89541390e-02  2.79992949e-02\n",
            " -1.36022521e-02 -4.32368740e-02  1.85830276e-02  7.63636417e-05\n",
            "  2.43514427e-03 -3.73319793e-03 -1.72279812e-02  1.01292422e-02\n",
            "  1.98437423e-02 -2.60018427e-02 -3.40176979e-03  1.09125171e-02\n",
            " -4.16363701e-02  3.37031744e-02 -2.81635392e-02  1.79126337e-02\n",
            " -4.53095250e-02 -1.09819863e-02 -2.20822985e-03  1.99103057e-02\n",
            "  3.56371962e-02 -3.11799254e-02  3.78751867e-02 -1.41409133e-02\n",
            " -2.16907021e-02  2.73019746e-02  3.69817764e-03  6.35388270e-02\n",
            "  1.22669144e-02 -6.02274062e-03 -7.60759786e-03 -1.86565295e-02\n",
            " -5.64718852e-03 -2.20049499e-03 -1.31825358e-02  1.67724490e-02\n",
            " -3.77264768e-02  2.97894888e-02 -5.01569919e-02  4.89087962e-02\n",
            " -6.07444458e-02 -8.39418694e-02 -5.09001613e-02  1.81767903e-02\n",
            "  6.66732788e-02 -3.30040907e-03 -2.82399380e-03 -5.35406843e-02\n",
            "  3.90341468e-02  2.19853055e-02  3.13555524e-02 -3.36527228e-02\n",
            "  1.96913332e-02  1.67883858e-02  5.04002981e-02  3.08058062e-03\n",
            " -7.24756741e-04  4.42907028e-02 -4.12959186e-03  4.29328792e-02\n",
            " -6.62552044e-02  1.16052642e-03 -2.81716529e-02  1.56886019e-02\n",
            "  9.78133380e-02  5.53594567e-02 -1.39378980e-02  2.12307293e-02\n",
            " -1.30956164e-02 -6.82028085e-02 -8.09035148e-04  4.99291979e-02\n",
            " -2.69265492e-02 -2.97803674e-02  3.84462215e-02  1.97353885e-02\n",
            "  3.37088443e-02  1.65873785e-02  5.77314151e-03 -3.04897372e-02\n",
            " -1.52511746e-02 -3.56159620e-02 -8.69219285e-03 -5.42296833e-33\n",
            "  3.24369315e-03 -3.46329585e-02  3.58933099e-02  1.83771104e-02\n",
            " -2.17504948e-02 -3.26411724e-02  2.88396631e-03  1.50463907e-02\n",
            " -1.75259868e-03 -1.99418906e-02 -6.10352168e-03  2.23847255e-02\n",
            " -8.78938241e-04  2.48685088e-02  3.39736044e-02  2.75592860e-02\n",
            "  3.37792151e-02  3.98564972e-02  2.55545750e-02  1.83042325e-02\n",
            " -2.92878635e-02  5.18083246e-03  8.37838568e-04 -3.66560556e-02\n",
            " -3.46733145e-02  3.82687114e-02  5.50822588e-03 -4.35187742e-02\n",
            "  2.44077630e-02  3.54167446e-02 -2.13442668e-02  2.86623556e-02\n",
            " -2.65410112e-04  3.73409986e-02 -8.68166331e-03  3.04787070e-03\n",
            " -2.71682125e-02 -3.85088287e-02 -6.12388104e-02 -2.00841762e-03\n",
            " -1.22080352e-02 -8.67197961e-02  3.75359156e-03 -1.77707188e-02\n",
            "  8.32484756e-03 -1.69167407e-02  7.02404007e-02  3.32233943e-02\n",
            "  4.34313081e-02  1.47017278e-02 -1.25546902e-01  1.50866257e-02\n",
            " -5.43164313e-02 -1.79148326e-03  4.99600992e-02 -1.53786335e-02\n",
            "  3.32683176e-02 -3.07708941e-02 -1.83896571e-02  9.45800170e-03\n",
            " -4.60291430e-02 -2.03868630e-03  2.62428429e-02 -5.00789136e-02\n",
            "  2.01836117e-02  6.08983040e-02 -2.01180391e-02 -2.60053724e-02\n",
            "  1.05925268e-02 -3.31153385e-02  1.62595455e-02  7.77862445e-02\n",
            " -1.90733070e-03 -5.62884938e-03  1.43716345e-02 -4.06833440e-02\n",
            " -5.14971055e-02  1.66212893e-04 -3.33052175e-03  1.44688589e-02\n",
            "  4.24896309e-04  3.04453429e-02 -1.83636565e-02  1.51191920e-03\n",
            "  2.99861431e-02 -3.68002206e-02  8.35622381e-03 -3.31025682e-02\n",
            "  2.66911536e-02  5.47830435e-03 -1.80524085e-02  2.42577959e-02\n",
            "  5.72708854e-03 -5.93372099e-02  1.04358397e-01 -9.87923332e-03\n",
            " -1.36106005e-02  5.79999052e-02  2.50108484e-02  2.89337076e-02\n",
            " -3.20520811e-02 -3.40233222e-02 -3.41698490e-02 -2.76982095e-02\n",
            "  6.47003055e-02  1.50798196e-02 -1.61925089e-02  3.03266384e-02\n",
            " -2.67188028e-02 -3.67774479e-02 -2.27845423e-02 -5.36433794e-02\n",
            "  1.90499797e-02 -3.42503227e-02  1.32688135e-02 -5.41317556e-03\n",
            "  7.49741495e-03 -7.36937160e-04 -3.08569428e-02  3.82288396e-02\n",
            " -2.08311286e-02 -3.43154892e-02  5.60240354e-03  1.44999241e-02\n",
            " -3.76364030e-02 -5.11783026e-02 -3.51075158e-02  1.71867851e-02\n",
            "  1.50721632e-02 -9.62026566e-02 -1.53544759e-02  1.58376191e-02\n",
            "  2.42940956e-07 -5.88800525e-03  7.68796578e-02  5.86063452e-02\n",
            "  2.21232865e-02 -2.36690883e-02  5.25273867e-02  1.48661695e-02\n",
            "  7.34176766e-03 -4.98924078e-03  4.37413976e-02 -1.28331855e-02\n",
            "  3.37342285e-02 -1.10815056e-02 -1.33937830e-02 -7.80064240e-02\n",
            " -1.36330752e-02  1.94749180e-02  1.91744871e-03 -3.00251674e-02\n",
            "  1.02716462e-04  9.54532623e-02  1.19654007e-01  3.73372361e-02\n",
            "  4.25128825e-03  2.05129981e-02 -3.85415033e-02 -1.90614313e-02\n",
            "  5.88792861e-02  6.81264400e-02 -3.12595516e-02 -6.50440454e-02\n",
            "  2.48044003e-02  3.90103814e-04  7.54762217e-02 -3.46075669e-02\n",
            "  1.32949557e-02  4.14005741e-02  3.07568815e-02  5.50352130e-03\n",
            " -1.53087813e-03  2.75993403e-02  6.46029925e-03  1.05398213e-02\n",
            " -3.09298541e-02  4.60232347e-02 -3.64920497e-02 -1.39539959e-02\n",
            " -3.53721119e-02  7.97802699e-04  1.40632764e-02  1.80258434e-02\n",
            " -1.43368859e-02  2.19216105e-03 -3.96873355e-02 -1.17282039e-02\n",
            " -4.45219912e-02  8.05771444e-03 -4.04862091e-02  3.56148668e-02\n",
            "  5.12853190e-02 -6.64038584e-02 -5.32594845e-02  8.92898999e-03\n",
            "  1.56423841e-02  1.02110982e-01  8.10765196e-03 -4.03859932e-03\n",
            "  2.02352651e-34 -1.38293868e-02 -1.17623918e-02  1.51005890e-02\n",
            "  8.25896114e-02  2.39228606e-02 -1.10378098e-02  3.65661248e-03\n",
            " -7.44783366e-03  2.94555184e-02  3.53000127e-03 -6.10421747e-02]\n",
            "\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "single_sentence = \"Yo! How cool are embeddings?\"\n",
        "single_embedding = embedding_model.encode(single_sentence)\n",
        "print(f\"Sentence: {single_sentence}\")\n",
        "print(f\"Embedding:\\n{single_embedding}\")\n",
        "print(f\"Embedding size: {single_embedding.shape}\")"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "IbN6V99bignw",
        "outputId": "c32514e5-2ea3-4699-e25c-7eecc2d7ffdf"
      },
      "execution_count": 23,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Sentence: Yo! How cool are embeddings?\n",
            "Embedding:\n",
            "[-1.97448116e-02 -4.51077055e-03 -4.98486962e-03  6.55444860e-02\n",
            " -9.87674389e-03  2.72836108e-02  3.66426110e-02 -3.30219767e-03\n",
            "  8.50078650e-03  8.24952498e-03 -2.28497703e-02  4.02430147e-02\n",
            " -5.75200692e-02  6.33691847e-02  4.43207137e-02 -4.49506715e-02\n",
            "  1.25284614e-02 -2.52011847e-02 -3.55293006e-02  1.29559003e-02\n",
            "  8.67021922e-03 -1.92917790e-02  3.55635840e-03  1.89505480e-02\n",
            " -1.47128161e-02 -9.39848833e-03  7.64175924e-03  9.62184742e-03\n",
            " -5.98920882e-03 -3.90168726e-02 -5.47824651e-02 -5.67456335e-03\n",
            "  1.11644426e-02  4.08067517e-02  1.76319088e-06  9.15305596e-03\n",
            " -8.77257995e-03  2.39382870e-02 -2.32784245e-02  8.04999843e-02\n",
            "  3.19176875e-02  5.12598455e-03 -1.47708450e-02 -1.62525177e-02\n",
            " -6.03213124e-02 -4.35689688e-02  4.51211594e-02 -1.79053694e-02\n",
            "  2.63366792e-02 -3.47866528e-02 -8.89172778e-03 -5.47675341e-02\n",
            " -1.24372439e-02 -2.38606706e-02  8.33496898e-02  5.71241677e-02\n",
            "  1.13328267e-02 -1.49595067e-02  9.20377970e-02  2.72710025e-02\n",
            " -1.42185586e-02  1.91208981e-02  1.49963303e-02 -3.12198866e-02\n",
            "  8.99579152e-02  4.51188162e-02  2.58020535e-02 -5.51739102e-03\n",
            "  1.15909716e-02  4.72100899e-02 -1.51019190e-02  1.70818325e-02\n",
            " -7.22596981e-03  3.45763564e-02 -8.76465999e-03  5.22016846e-02\n",
            " -6.49008900e-02 -4.31378335e-02  6.36964813e-02  4.02881242e-02\n",
            " -1.99042782e-02  5.39059704e-03  1.28820529e-02 -4.81255502e-02\n",
            "  4.58800681e-02 -2.17094608e-02  1.89203694e-02 -3.46344374e-02\n",
            " -1.66456699e-02  7.65162846e-03 -2.26693675e-02 -1.96454227e-02\n",
            "  1.87632181e-02  1.01383040e-02  6.85412809e-02 -5.39850304e-03\n",
            " -3.38225579e-03  4.08412963e-02  4.98623997e-02 -1.16485534e-02\n",
            "  8.91738907e-02  4.02785838e-02 -3.64719215e-03  4.37758975e-02\n",
            " -2.96080597e-02 -5.53753832e-03 -2.00208891e-02 -2.01982595e-02\n",
            "  4.59849164e-02  2.29337290e-02 -5.37306331e-02 -3.19279172e-02\n",
            "  1.37536845e-03  6.25036359e-02 -2.18308680e-02 -6.43255934e-02\n",
            " -2.24791840e-02  3.31955217e-02 -3.12837772e-02  5.17936423e-02\n",
            " -2.84003261e-02  2.55067777e-02  3.36493775e-02  7.50668272e-02\n",
            " -4.46472317e-03 -4.87705469e-02 -7.35218897e-02 -5.46353199e-02\n",
            "  8.88524856e-03  2.95797866e-02 -9.95699130e-03 -6.32761279e-03\n",
            "  4.46259752e-02 -1.58483982e-02 -1.71330497e-02  3.36603001e-02\n",
            " -1.57665915e-03 -5.45971878e-02  2.91160922e-02 -2.80596036e-02\n",
            "  2.96793506e-02  5.12153134e-02  1.48765966e-02 -4.76487912e-02\n",
            "  1.26052313e-02  1.49851339e-03  1.33206481e-02 -2.82469429e-02\n",
            " -3.29254642e-02 -8.53569247e-03 -5.27607873e-02  7.29350299e-02\n",
            " -6.41821176e-02 -2.51785177e-03 -9.02637839e-03 -1.10469048e-03\n",
            "  1.57514382e-02  4.30823267e-02  1.12269474e-02 -3.54585350e-02\n",
            "  4.95163240e-02  1.21270753e-02  1.66343944e-03 -5.06922835e-03\n",
            " -1.11001981e-02 -8.66925716e-03 -3.26440036e-02 -3.98021601e-02\n",
            " -2.05970593e-02  1.09074460e-02 -6.62528276e-02  3.71707082e-02\n",
            " -3.74916568e-02 -3.24730650e-02  5.85900694e-02  8.48081112e-02\n",
            "  3.92413139e-02  3.15815173e-02  3.78386192e-02 -1.35472463e-02\n",
            "  5.95062524e-02  2.58904770e-02 -1.31900040e-02  6.30589649e-02\n",
            "  3.27136293e-02  6.92142686e-03 -1.42607335e-02  7.76676536e-02\n",
            " -1.16102919e-02 -3.66428010e-02 -2.83837840e-02  2.72279773e-02\n",
            "  2.49364488e-02 -4.22296487e-03 -3.63100693e-02 -2.04887725e-02\n",
            "  3.98861170e-02 -2.64726747e-02  4.41380590e-03 -5.19635379e-02\n",
            "  1.71031425e-05  4.81284857e-02  2.04450656e-02  9.84970033e-02\n",
            "  3.68267260e-02  1.53404567e-02  7.50978128e-04 -3.38638499e-02\n",
            " -2.69872826e-02  4.72444631e-02  4.56701554e-02 -3.49246487e-02\n",
            " -1.18770609e-02  3.45576694e-03 -6.32300740e-03 -4.78412546e-02\n",
            "  1.84098631e-02 -2.23157462e-02 -3.70727777e-02  5.87339625e-02\n",
            "  6.22731261e-03 -1.46716302e-02  7.29222819e-02  2.21960037e-03\n",
            " -6.53120056e-02  3.51679660e-02 -1.54901426e-02  6.01420999e-02\n",
            " -9.41001624e-03  2.81196851e-02 -1.12652015e-02  5.24159812e-04\n",
            "  1.01888627e-01 -5.69957457e-02 -3.52360681e-02 -5.20478189e-03\n",
            " -8.46638530e-03  1.39209256e-02  1.80780459e-02 -1.10493951e-01\n",
            "  5.13116606e-02 -4.36432101e-02  2.84142829e-02  9.55936499e-03\n",
            "  4.28096876e-02 -3.95833775e-02  5.25829196e-02  1.92814786e-02\n",
            "  3.44885583e-03 -1.76871512e-02  3.85699905e-02  6.92503573e-03\n",
            " -3.59440297e-02 -2.63612997e-02 -4.96697752e-03  4.24525589e-02\n",
            " -4.22464013e-02  3.45900049e-03  3.55866738e-02 -1.68201923e-02\n",
            "  3.54331844e-02  4.52805730e-03  6.46290975e-03 -2.17710938e-02\n",
            " -2.50032954e-02  1.41418697e-02  1.51257757e-02 -2.99570095e-02\n",
            " -3.94227579e-02  1.87821817e-02 -1.68784207e-03 -9.83500038e-04\n",
            " -3.26322019e-02  5.06563019e-03 -8.90461728e-03 -1.55095393e-02\n",
            "  1.87758375e-02 -4.52473201e-02 -1.72957368e-02  3.50973345e-02\n",
            "  1.33018205e-02  1.00633139e-02 -4.36593257e-02 -1.68618988e-02\n",
            " -1.91048421e-02  6.35489821e-02  8.08325410e-03 -1.02532730e-02\n",
            " -4.53626877e-03 -4.60835695e-02 -1.64704248e-02 -6.24686107e-03\n",
            "  2.58868579e-02 -6.39063939e-02 -7.82401394e-03 -2.36076415e-02\n",
            "  2.74617691e-02  5.80534860e-02 -3.40748765e-02  6.46012202e-02\n",
            "  2.00062506e-02 -2.14934926e-02  1.69361029e-02 -5.54070575e-03\n",
            " -2.40397546e-02  3.09444238e-02 -2.34439666e-03  3.02589443e-02\n",
            " -4.57217321e-02  2.00641155e-02 -2.57117227e-02 -1.13377406e-03\n",
            " -3.57524604e-02  6.92953467e-02  2.80848774e-03  3.58742177e-02\n",
            " -1.52722159e-02 -3.41523737e-02  1.80923473e-02  1.65400244e-02\n",
            "  1.31705850e-02 -6.36670925e-03  5.49302921e-02  8.47313832e-03\n",
            " -4.26077433e-02  2.17084587e-02 -4.89023477e-02  1.18851785e-04\n",
            "  5.16287647e-02  7.59220857e-04  1.59222875e-02 -1.61299706e-02\n",
            "  1.44981248e-02  2.19509210e-02  3.02651562e-02 -3.44151333e-02\n",
            " -4.80380096e-02 -3.71689945e-02  4.68194932e-02 -3.46219353e-02\n",
            "  4.74861532e-04 -4.34820950e-02 -1.80125143e-02 -6.44844398e-02\n",
            " -2.66967565e-02  3.63660902e-02 -3.76219675e-02 -1.64600983e-02\n",
            "  2.20248327e-02  2.16080312e-04  3.64510454e-02 -2.76135765e-02\n",
            " -5.87056717e-03  6.97318371e-03 -1.25863450e-03  2.12774165e-02\n",
            "  6.21470995e-03 -3.57214659e-02 -5.09366281e-02  2.85553224e-02\n",
            "  6.48822486e-02  3.45731415e-02 -2.57281270e-02  4.52561444e-03\n",
            " -4.63605747e-02  3.32225338e-02  1.69973925e-03 -1.29355500e-02\n",
            " -3.03734131e-02  1.23608978e-02 -3.17957747e-04  1.66231990e-02\n",
            "  1.04892189e-02  1.71540752e-02  1.88860409e-02 -3.62256467e-02\n",
            "  4.65737768e-02  2.17128210e-02  4.97535989e-02  3.03067546e-02\n",
            "  1.59914012e-03 -6.30024746e-02 -6.34594914e-03  1.07304622e-04\n",
            " -5.56749525e-03  2.37264968e-02 -8.38230271e-03  5.38897254e-02\n",
            " -9.08977017e-02 -1.37359872e-02  1.18454583e-02  3.37257539e-03\n",
            " -2.81855483e-02  1.56333228e-03  2.22415458e-02  6.21023960e-02\n",
            " -8.68119821e-02 -4.40635020e-03 -1.63995549e-02  1.69665273e-02\n",
            " -1.34548452e-02  3.00696120e-03 -2.14618240e-02 -2.09503789e-02\n",
            " -1.39877591e-02 -1.23850694e-02  5.39979152e-02  6.03614561e-02\n",
            "  2.52982210e-02 -1.29661456e-01 -1.08081467e-01 -4.15774900e-03\n",
            " -7.20272539e-03  2.75885276e-02  4.87622321e-02 -2.95970980e-02\n",
            " -4.32555042e-02  2.75215656e-02  1.35718873e-02  3.87699716e-02\n",
            "  2.42039673e-02 -2.70842165e-02  8.59166235e-02 -1.98402982e-02\n",
            " -2.26344205e-02 -6.24927804e-02 -1.56845022e-02  4.11566831e-02\n",
            "  1.66952405e-02  7.97291845e-02 -3.24839242e-02  1.46561884e-03\n",
            " -3.27906273e-02  6.44815490e-02  2.09739413e-02 -7.56986067e-02\n",
            " -1.31789874e-03  2.24045897e-03 -6.60841586e-03 -6.84251785e-02\n",
            " -5.36860572e-03  6.55136183e-02  5.45558985e-03  1.58993863e-02\n",
            " -2.18052231e-02  2.16424419e-03  4.72954288e-03 -7.25655705e-02\n",
            " -1.59349740e-02 -1.05623305e-02  2.70786788e-02  1.93766574e-03\n",
            " -4.45123315e-02  2.89783385e-02  2.43084226e-02 -1.73885785e-02\n",
            " -3.80669758e-02 -3.06799151e-02 -3.24778259e-02 -4.33339184e-04\n",
            "  2.55846083e-02  2.70478483e-02 -1.72519751e-04 -4.93688567e-04\n",
            " -7.12094754e-02  5.69768772e-02  6.61400855e-02 -3.87268104e-02\n",
            " -1.30683519e-02  1.00663211e-02 -2.17740331e-02  1.92212239e-02\n",
            "  7.66691146e-03 -3.86652574e-02 -1.66108645e-02 -3.41467746e-02\n",
            " -1.04185315e-02  1.75906308e-02 -1.23776542e-02 -1.68433841e-02\n",
            " -2.40113102e-02  4.70198505e-03  4.88461461e-03  4.73663509e-02\n",
            "  4.34111953e-02 -8.08343571e-03 -2.48272233e-02 -1.93978269e-02\n",
            " -3.16525102e-02 -1.56419035e-02 -7.79948430e-03  1.28888441e-02\n",
            "  2.61943396e-02  3.65819340e-03  5.79228774e-02  5.43151647e-02\n",
            " -5.05586788e-02  1.78927090e-03 -2.45471150e-02 -2.47595999e-02\n",
            "  3.60356877e-03  1.94152240e-02 -4.23822924e-02 -1.86907277e-02\n",
            "  2.32945494e-02 -3.17982435e-02 -2.60645691e-02 -5.40362298e-03\n",
            " -3.82069871e-02  2.21719909e-02  1.33360280e-02  5.58054484e-02\n",
            " -3.57716866e-02 -3.54791656e-02  1.27723301e-02  6.80177063e-02\n",
            "  5.37152812e-02  2.54151784e-02 -1.16638504e-02 -1.07512977e-02\n",
            " -9.74432100e-03  7.20507465e-03  9.21899080e-03 -4.73684780e-02\n",
            " -3.89389903e-03  3.11453529e-02  3.62331979e-02 -1.65902935e-02\n",
            " -3.63394991e-02  1.95634607e-02 -2.15058494e-02 -7.04767322e-03\n",
            "  9.13175754e-03 -4.05358896e-02 -3.67076248e-02  1.16995938e-01\n",
            "  1.17913887e-01  8.00502896e-02 -1.61983129e-02 -2.00733747e-02\n",
            " -5.54062016e-02 -7.22410306e-02  2.14558002e-02 -4.46405145e-04\n",
            " -1.42903179e-02  8.35540146e-03  3.34207676e-02  1.42891202e-02\n",
            "  4.62512635e-02 -3.53420191e-02  1.68673266e-02 -2.99732201e-03\n",
            " -5.44997007e-02 -4.80719730e-02  9.47561057e-04 -6.29721507e-33\n",
            " -2.30286438e-02 -2.51128357e-02 -5.35218976e-02 -2.09470019e-02\n",
            " -6.79717492e-03 -4.64015529e-02 -4.49633691e-03  1.65114887e-02\n",
            " -1.12677794e-02  1.33013809e-02 -1.72552820e-02 -1.96653586e-02\n",
            "  5.53236343e-03  1.02775376e-02  9.47373686e-04 -2.24022288e-02\n",
            "  5.30364849e-02  7.77775934e-03 -9.48474091e-03  2.25515869e-02\n",
            " -4.34495928e-03  2.25208756e-02  1.98086575e-02 -7.57428482e-02\n",
            " -4.36680904e-03  2.50829905e-02  2.59393696e-02 -3.07077561e-02\n",
            "  7.04764351e-02  8.63500684e-02 -7.75880739e-02  1.59991533e-02\n",
            " -5.04692420e-02  4.88355197e-02  3.74992355e-03 -6.12926669e-04\n",
            " -3.87277901e-02 -2.32235119e-02 -3.63983437e-02 -5.07016154e-03\n",
            "  1.10517768e-02 -3.26515660e-02  3.68386693e-02 -4.54948992e-02\n",
            " -1.18526642e-03  1.92688184e-03  2.18783580e-02  2.71092802e-02\n",
            " -4.06264216e-02  6.99159801e-02 -7.33282045e-02 -8.15295149e-03\n",
            " -1.42555609e-02  3.78038781e-03  1.20974362e-01 -6.68213218e-02\n",
            "  3.05051152e-02  1.24480613e-02 -4.59293984e-02  1.03873098e-02\n",
            " -3.97978052e-02 -1.33042196e-02 -1.59402415e-02 -4.29347157e-02\n",
            "  4.05275710e-02  7.07073808e-02 -4.50928546e-02 -3.62474471e-02\n",
            " -1.87588781e-02  1.60928648e-02  2.12657135e-02  6.70025423e-02\n",
            "  3.25864181e-02  1.51125938e-02  3.20371501e-02 -1.35436170e-02\n",
            "  2.31779981e-02 -1.13125844e-02  1.23795634e-02 -3.73516977e-02\n",
            "  1.55545224e-03  2.15824582e-02 -3.49442437e-02 -2.97690742e-02\n",
            "  2.32397411e-02 -1.25703337e-02 -1.09432973e-02 -8.87968689e-02\n",
            " -2.20183991e-02 -1.18423048e-02 -5.71083613e-02  3.91810127e-02\n",
            " -1.98827125e-02 -5.59270494e-02  7.60342227e-03  2.23641824e-02\n",
            " -1.86267253e-02  3.79805267e-02 -8.79652449e-04 -5.26199490e-02\n",
            "  2.05068011e-03  1.72814839e-02 -4.84028980e-02 -1.87740419e-02\n",
            "  1.91171608e-07  2.06594449e-02  3.03575210e-02  5.71851479e-03\n",
            " -5.33938222e-02 -2.46520881e-02  1.80340782e-02 -3.39978002e-02\n",
            "  3.46113593e-05 -6.40035793e-02  2.50049569e-02 -2.04111841e-02\n",
            " -2.72037974e-03 -3.55961695e-02  2.71923114e-02  6.48940206e-02\n",
            "  9.83432983e-04 -4.38490510e-02 -4.45297174e-02 -7.44889490e-03\n",
            "  1.15205254e-02 -2.91254069e-03 -2.15494987e-02  2.84254062e-03\n",
            "  4.29398939e-02 -6.09041229e-02 -7.86320306e-03 -3.90126230e-03\n",
            "  2.47718305e-07  7.75886059e-04  7.47737065e-02  1.99318118e-03\n",
            " -6.07220875e-03  3.69210728e-02  2.78421324e-02 -5.64900041e-02\n",
            "  1.61058903e-02 -9.50827170e-03 -2.60848785e-03 -2.45737229e-02\n",
            "  1.91390775e-02  5.08195609e-02  2.61258744e-02 -1.03838101e-01\n",
            " -3.05815265e-02 -3.53345163e-02 -4.07038108e-02 -2.19843015e-02\n",
            " -2.24092193e-02  5.05567454e-02  7.22607598e-02  5.54789975e-02\n",
            "  4.89434041e-02  3.37949395e-03 -6.84760064e-02  7.10320659e-03\n",
            "  3.15671461e-03  4.78091799e-02 -7.19795078e-02 -3.30301970e-02\n",
            "  3.19158584e-02  1.76431192e-03 -4.62790132e-02 -1.96378492e-02\n",
            "  1.67493820e-02  4.73603830e-02 -2.09441632e-02  7.10244663e-03\n",
            "  4.53146063e-02 -4.76523638e-02 -4.74881381e-02  1.00799864e-02\n",
            " -8.39594007e-02  3.36930230e-02 -3.72189060e-02  1.19432965e-02\n",
            " -3.16896290e-02 -1.29724119e-03 -1.55541236e-02  1.81727745e-02\n",
            " -1.49368513e-02 -1.70671456e-02 -4.19716686e-02  3.94660467e-03\n",
            " -2.24302057e-02  2.07292736e-02 -4.61415686e-02  8.50217510e-03\n",
            " -2.56864522e-02 -1.65337678e-02 -1.51846372e-02 -1.00041572e-02\n",
            "  2.19642837e-02  2.61104815e-02  7.31359124e-02 -1.83709543e-02\n",
            "  1.93979481e-34 -7.27934530e-03  5.96488127e-03  4.44310270e-02\n",
            "  4.14822660e-02  1.12916902e-02 -1.93217508e-02  4.41879369e-02\n",
            " -8.93792044e-03  3.61120030e-02 -5.52125759e-02 -2.89572161e-02]\n",
            "Embedding size: (768,)\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "%%time\n",
        "\n",
        "# Send the model to the GPU\n",
        "embedding_model.to(\"cuda\") # requires a GPU installed, for reference on my local machine, I'm using a NVIDIA RTX 4090\n",
        "\n",
        "# Create embeddings one by one on the GPU\n",
        "for item in tqdm(pages_and_chunks_over_min_token_len):\n",
        "    item[\"embedding\"] = embedding_model.encode(item[\"sentence_chunk\"])"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 86,
          "referenced_widgets": [
            "66aa4dd8cfe040ad96d92733e55ad1c5",
            "0250334a34dd41a38a7135ca6da6c234",
            "ce073b1d172543b4aab5950ebfdf1459",
            "4dec3d51a88e483f8bdb84d57c30bf70",
            "aa002c60e0a44244af89f6c13ffb718f",
            "6f573805b7ee49aca5d6e06670b585e6",
            "17e995d86e68426b93147a7f1430700c",
            "5cb0d781db454b59bf5ca34cea5fb991",
            "57b442944ac346758c164c037d80dcd5",
            "670637917949438082271350428bbd5c",
            "4ed404f08720460f965e5bf773e4919d"
          ]
        },
        "id": "K6rkT2sUjMsH",
        "outputId": "1e22a038-03c8-4d0f-faca-5a264705ed95"
      },
      "execution_count": 24,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "  0%|          | 0/35 [00:00<?, ?it/s]"
            ],
            "application/vnd.jupyter.widget-view+json": {
              "version_major": 2,
              "version_minor": 0,
              "model_id": "66aa4dd8cfe040ad96d92733e55ad1c5"
            }
          },
          "metadata": {}
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "CPU times: user 3.1 s, sys: 338 ms, total: 3.44 s\n",
            "Wall time: 2.49 s\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# Turn text chunks into a single list\n",
        "text_chunks = [item[\"sentence_chunk\"] for item in pages_and_chunks_over_min_token_len]"
      ],
      "metadata": {
        "id": "33tgBb-8jk6K"
      },
      "execution_count": 25,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "%%time\n",
        "\n",
        "# Embed all texts in batches\n",
        "text_chunk_embeddings = embedding_model.encode(text_chunks,\n",
        "                                               batch_size=32, # you can use different batch sizes here for speed/performance, I found 32 works well for this use case\n",
        "                                               convert_to_tensor=True) # optional to return embeddings as tensor instead of array\n",
        "\n",
        "text_chunk_embeddings"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "-dKqBZdfjqZH",
        "outputId": "5e28bd4f-b9eb-494a-8871-6659bcf04e0b"
      },
      "execution_count": 26,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "CPU times: user 940 ms, sys: 4.46 ms, total: 944 ms\n",
            "Wall time: 818 ms\n"
          ]
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "tensor([[-0.0199,  0.0228,  0.0031,  ..., -0.0688,  0.0063,  0.0173],\n",
              "        [-0.0090,  0.0336, -0.0204,  ..., -0.0468, -0.0234, -0.0376],\n",
              "        [-0.0414,  0.0438, -0.0120,  ..., -0.0430, -0.0261, -0.0288],\n",
              "        ...,\n",
              "        [-0.0407, -0.0263, -0.0056,  ..., -0.0326, -0.0325, -0.0200],\n",
              "        [-0.0252,  0.0127, -0.0063,  ..., -0.0500, -0.0211, -0.0151],\n",
              "        [ 0.0117, -0.0427,  0.0029,  ..., -0.0719,  0.0091,  0.0033]],\n",
              "       device='cuda:0')"
            ]
          },
          "metadata": {},
          "execution_count": 26
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# Save embeddings to file\n",
        "text_chunks_and_embeddings_df = pd.DataFrame(pages_and_chunks_over_min_token_len)\n",
        "embeddings_df_save_path = \"text_chunks_and_embeddings_df.csv\"\n",
        "text_chunks_and_embeddings_df.to_csv(embeddings_df_save_path, index=False)"
      ],
      "metadata": {
        "id": "qebSzOXcjwus"
      },
      "execution_count": 27,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "# Import saved file and view\n",
        "text_chunks_and_embedding_df_load = pd.read_csv(embeddings_df_save_path)\n",
        "text_chunks_and_embedding_df_load.head()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 379
        },
        "id": "iIfMj5qij4-a",
        "outputId": "e12f3dfd-6d22-4656-b4ff-542ff9a87951"
      },
      "execution_count": 28,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "   page_number                                     sentence_chunk  \\\n",
              "0          -41  REAL-TIME SEMANTIC SEGMENTATION ON FPGAS FOR A...   \n",
              "1          -40  challenges faced by, for example, the automoti...   \n",
              "2          -40  By applying aggressive ﬁlter-reduction and qua...   \n",
              "3          -40  With these steps, we obtain a good balance bet...   \n",
              "4          -40  In this way all inputs are smaller than one an...   \n",
              "\n",
              "   chunk_char_count  chunk_word_count  chunk_token_count  \\\n",
              "0              2364               329             591.00   \n",
              "1              1816               256             454.00   \n",
              "2              1633               234             408.25   \n",
              "3              1188               190             297.00   \n",
              "4               707               129             176.75   \n",
              "\n",
              "                                           embedding  \n",
              "0  [-1.99189372e-02  2.27944255e-02  3.12705827e-...  \n",
              "1  [-9.01952107e-03  3.35726589e-02 -2.03959588e-...  \n",
              "2  [-4.13998924e-02  4.38373275e-02 -1.20260157e-...  \n",
              "3  [-9.95968096e-03  4.90480065e-02  3.64915580e-...  \n",
              "4  [-6.82296976e-02 -3.08206026e-02  1.60923079e-...  "
            ],
            "text/html": [
              "\n",
              "  <div id=\"df-d55971a6-c8ea-41dc-8554-4a2cf9990d92\" class=\"colab-df-container\">\n",
              "    <div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>page_number</th>\n",
              "      <th>sentence_chunk</th>\n",
              "      <th>chunk_char_count</th>\n",
              "      <th>chunk_word_count</th>\n",
              "      <th>chunk_token_count</th>\n",
              "      <th>embedding</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>-41</td>\n",
              "      <td>REAL-TIME SEMANTIC SEGMENTATION ON FPGAS FOR A...</td>\n",
              "      <td>2364</td>\n",
              "      <td>329</td>\n",
              "      <td>591.00</td>\n",
              "      <td>[-1.99189372e-02  2.27944255e-02  3.12705827e-...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>-40</td>\n",
              "      <td>challenges faced by, for example, the automoti...</td>\n",
              "      <td>1816</td>\n",
              "      <td>256</td>\n",
              "      <td>454.00</td>\n",
              "      <td>[-9.01952107e-03  3.35726589e-02 -2.03959588e-...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>-40</td>\n",
              "      <td>By applying aggressive ﬁlter-reduction and qua...</td>\n",
              "      <td>1633</td>\n",
              "      <td>234</td>\n",
              "      <td>408.25</td>\n",
              "      <td>[-4.13998924e-02  4.38373275e-02 -1.20260157e-...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>-40</td>\n",
              "      <td>With these steps, we obtain a good balance bet...</td>\n",
              "      <td>1188</td>\n",
              "      <td>190</td>\n",
              "      <td>297.00</td>\n",
              "      <td>[-9.95968096e-03  4.90480065e-02  3.64915580e-...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>-40</td>\n",
              "      <td>In this way all inputs are smaller than one an...</td>\n",
              "      <td>707</td>\n",
              "      <td>129</td>\n",
              "      <td>176.75</td>\n",
              "      <td>[-6.82296976e-02 -3.08206026e-02  1.60923079e-...</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>\n",
              "    <div class=\"colab-df-buttons\">\n",
              "\n",
              "  <div class=\"colab-df-container\">\n",
              "    <button class=\"colab-df-convert\" onclick=\"convertToInteractive('df-d55971a6-c8ea-41dc-8554-4a2cf9990d92')\"\n",
              "            title=\"Convert this dataframe to an interactive table.\"\n",
              "            style=\"display:none;\">\n",
              "\n",
              "  <svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\" viewBox=\"0 -960 960 960\">\n",
              "    <path d=\"M120-120v-720h720v720H120Zm60-500h600v-160H180v160Zm220 220h160v-160H400v160Zm0 220h160v-160H400v160ZM180-400h160v-160H180v160Zm440 0h160v-160H620v160ZM180-180h160v-160H180v160Zm440 0h160v-160H620v160Z\"/>\n",
              "  </svg>\n",
              "    </button>\n",
              "\n",
              "  <style>\n",
              "    .colab-df-container {\n",
              "      display:flex;\n",
              "      gap: 12px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert {\n",
              "      background-color: #E8F0FE;\n",
              "      border: none;\n",
              "      border-radius: 50%;\n",
              "      cursor: pointer;\n",
              "      display: none;\n",
              "      fill: #1967D2;\n",
              "      height: 32px;\n",
              "      padding: 0 0 0 0;\n",
              "      width: 32px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert:hover {\n",
              "      background-color: #E2EBFA;\n",
              "      box-shadow: 0px 1px 2px rgba(60, 64, 67, 0.3), 0px 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
              "      fill: #174EA6;\n",
              "    }\n",
              "\n",
              "    .colab-df-buttons div {\n",
              "      margin-bottom: 4px;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert {\n",
              "      background-color: #3B4455;\n",
              "      fill: #D2E3FC;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert:hover {\n",
              "      background-color: #434B5C;\n",
              "      box-shadow: 0px 1px 3px 1px rgba(0, 0, 0, 0.15);\n",
              "      filter: drop-shadow(0px 1px 2px rgba(0, 0, 0, 0.3));\n",
              "      fill: #FFFFFF;\n",
              "    }\n",
              "  </style>\n",
              "\n",
              "    <script>\n",
              "      const buttonEl =\n",
              "        document.querySelector('#df-d55971a6-c8ea-41dc-8554-4a2cf9990d92 button.colab-df-convert');\n",
              "      buttonEl.style.display =\n",
              "        google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
              "\n",
              "      async function convertToInteractive(key) {\n",
              "        const element = document.querySelector('#df-d55971a6-c8ea-41dc-8554-4a2cf9990d92');\n",
              "        const dataTable =\n",
              "          await google.colab.kernel.invokeFunction('convertToInteractive',\n",
              "                                                    [key], {});\n",
              "        if (!dataTable) return;\n",
              "\n",
              "        const docLinkHtml = 'Like what you see? Visit the ' +\n",
              "          '<a target=\"_blank\" href=https://colab.research.google.com/notebooks/data_table.ipynb>data table notebook</a>'\n",
              "          + ' to learn more about interactive tables.';\n",
              "        element.innerHTML = '';\n",
              "        dataTable['output_type'] = 'display_data';\n",
              "        await google.colab.output.renderOutput(dataTable, element);\n",
              "        const docLink = document.createElement('div');\n",
              "        docLink.innerHTML = docLinkHtml;\n",
              "        element.appendChild(docLink);\n",
              "      }\n",
              "    </script>\n",
              "  </div>\n",
              "\n",
              "\n",
              "<div id=\"df-f97f1ae3-6e50-49fd-aa10-62a487079851\">\n",
              "  <button class=\"colab-df-quickchart\" onclick=\"quickchart('df-f97f1ae3-6e50-49fd-aa10-62a487079851')\"\n",
              "            title=\"Suggest charts\"\n",
              "            style=\"display:none;\">\n",
              "\n",
              "<svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\"viewBox=\"0 0 24 24\"\n",
              "     width=\"24px\">\n",
              "    <g>\n",
              "        <path d=\"M19 3H5c-1.1 0-2 .9-2 2v14c0 1.1.9 2 2 2h14c1.1 0 2-.9 2-2V5c0-1.1-.9-2-2-2zM9 17H7v-7h2v7zm4 0h-2V7h2v10zm4 0h-2v-4h2v4z\"/>\n",
              "    </g>\n",
              "</svg>\n",
              "  </button>\n",
              "\n",
              "<style>\n",
              "  .colab-df-quickchart {\n",
              "      --bg-color: #E8F0FE;\n",
              "      --fill-color: #1967D2;\n",
              "      --hover-bg-color: #E2EBFA;\n",
              "      --hover-fill-color: #174EA6;\n",
              "      --disabled-fill-color: #AAA;\n",
              "      --disabled-bg-color: #DDD;\n",
              "  }\n",
              "\n",
              "  [theme=dark] .colab-df-quickchart {\n",
              "      --bg-color: #3B4455;\n",
              "      --fill-color: #D2E3FC;\n",
              "      --hover-bg-color: #434B5C;\n",
              "      --hover-fill-color: #FFFFFF;\n",
              "      --disabled-bg-color: #3B4455;\n",
              "      --disabled-fill-color: #666;\n",
              "  }\n",
              "\n",
              "  .colab-df-quickchart {\n",
              "    background-color: var(--bg-color);\n",
              "    border: none;\n",
              "    border-radius: 50%;\n",
              "    cursor: pointer;\n",
              "    display: none;\n",
              "    fill: var(--fill-color);\n",
              "    height: 32px;\n",
              "    padding: 0;\n",
              "    width: 32px;\n",
              "  }\n",
              "\n",
              "  .colab-df-quickchart:hover {\n",
              "    background-color: var(--hover-bg-color);\n",
              "    box-shadow: 0 1px 2px rgba(60, 64, 67, 0.3), 0 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
              "    fill: var(--button-hover-fill-color);\n",
              "  }\n",
              "\n",
              "  .colab-df-quickchart-complete:disabled,\n",
              "  .colab-df-quickchart-complete:disabled:hover {\n",
              "    background-color: var(--disabled-bg-color);\n",
              "    fill: var(--disabled-fill-color);\n",
              "    box-shadow: none;\n",
              "  }\n",
              "\n",
              "  .colab-df-spinner {\n",
              "    border: 2px solid var(--fill-color);\n",
              "    border-color: transparent;\n",
              "    border-bottom-color: var(--fill-color);\n",
              "    animation:\n",
              "      spin 1s steps(1) infinite;\n",
              "  }\n",
              "\n",
              "  @keyframes spin {\n",
              "    0% {\n",
              "      border-color: transparent;\n",
              "      border-bottom-color: var(--fill-color);\n",
              "      border-left-color: var(--fill-color);\n",
              "    }\n",
              "    20% {\n",
              "      border-color: transparent;\n",
              "      border-left-color: var(--fill-color);\n",
              "      border-top-color: var(--fill-color);\n",
              "    }\n",
              "    30% {\n",
              "      border-color: transparent;\n",
              "      border-left-color: var(--fill-color);\n",
              "      border-top-color: var(--fill-color);\n",
              "      border-right-color: var(--fill-color);\n",
              "    }\n",
              "    40% {\n",
              "      border-color: transparent;\n",
              "      border-right-color: var(--fill-color);\n",
              "      border-top-color: var(--fill-color);\n",
              "    }\n",
              "    60% {\n",
              "      border-color: transparent;\n",
              "      border-right-color: var(--fill-color);\n",
              "    }\n",
              "    80% {\n",
              "      border-color: transparent;\n",
              "      border-right-color: var(--fill-color);\n",
              "      border-bottom-color: var(--fill-color);\n",
              "    }\n",
              "    90% {\n",
              "      border-color: transparent;\n",
              "      border-bottom-color: var(--fill-color);\n",
              "    }\n",
              "  }\n",
              "</style>\n",
              "\n",
              "  <script>\n",
              "    async function quickchart(key) {\n",
              "      const quickchartButtonEl =\n",
              "        document.querySelector('#' + key + ' button');\n",
              "      quickchartButtonEl.disabled = true;  // To prevent multiple clicks.\n",
              "      quickchartButtonEl.classList.add('colab-df-spinner');\n",
              "      try {\n",
              "        const charts = await google.colab.kernel.invokeFunction(\n",
              "            'suggestCharts', [key], {});\n",
              "      } catch (error) {\n",
              "        console.error('Error during call to suggestCharts:', error);\n",
              "      }\n",
              "      quickchartButtonEl.classList.remove('colab-df-spinner');\n",
              "      quickchartButtonEl.classList.add('colab-df-quickchart-complete');\n",
              "    }\n",
              "    (() => {\n",
              "      let quickchartButtonEl =\n",
              "        document.querySelector('#df-f97f1ae3-6e50-49fd-aa10-62a487079851 button');\n",
              "      quickchartButtonEl.style.display =\n",
              "        google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
              "    })();\n",
              "  </script>\n",
              "</div>\n",
              "\n",
              "    </div>\n",
              "  </div>\n"
            ],
            "application/vnd.google.colaboratory.intrinsic+json": {
              "type": "dataframe",
              "variable_name": "text_chunks_and_embedding_df_load",
              "summary": "{\n  \"name\": \"text_chunks_and_embedding_df_load\",\n  \"rows\": 35,\n  \"fields\": [\n    {\n      \"column\": \"page_number\",\n      \"properties\": {\n        \"dtype\": \"number\",\n        \"std\": 3,\n        \"min\": -41,\n        \"max\": -31,\n        \"num_unique_values\": 11,\n        \"samples\": [\n          -36,\n          -41,\n          -32\n        ],\n        \"semantic_type\": \"\",\n        \"description\": \"\"\n      }\n    },\n    {\n      \"column\": \"sentence_chunk\",\n      \"properties\": {\n        \"dtype\": \"string\",\n        \"num_unique_values\": 35,\n        \"samples\": [\n          \"M. P., M. R., S. S., and V. L. are supported by the European Research Council (ERC) under the European Union\\u2019s Horizon 2020 research and innovation program (Grant Agreement No.772369). M. P. and N. G. are supported by the European Research Council (ERC) under the European Union\\u2019s Horizon 2020 research and innovation program (Grant Agreement No.966696).10\",\n          \"In our case, we use AUTOQKERAS [12]. In AUTOQKERAS, a hyperparameter search over individual layer quantization conditions and \\ufb01lter counts is performed. Since the model contains skip connections, the scan over number of \\ufb01lters needs to be handled with care. In particular, we use the block features of AUTOQKERAS to ensure that the \\ufb01lter count matches throughout a bottleneck, so that the tensor addition of the skip connection will have valid dimensions. The search for best hyperparameters, including the choice of indivdual quantizers for kernels and activations, is carried out using a Bayesian strategy where the balance between accuracy and resource usage is controlled by targeting a metric derived from them both[12]. In our search we permit e.g. a 4% decrease in accuracy if the resource usage also is halved at the same time.6\",\n          \"Including these, the average time per image increases to 720 ms.6 Conclusions In this paper, we demonstrate that we can perform semantic segmentation on a single FPGA on a Zynq MPSoC device using a compressed version of ENet. The network is compressed using automatic heterogeneous quantization at training time and a \\ufb01lter ablation procedure, and is then evaluated on the Cityscapes dataset. Inference is executed on hardware with a latency of 4.9 ms per image, utilizing 18% of the DSPs, 30% of the LUTs, 16% of the FFs and 25 % of the BRAMs. Processing the images in batches of ten results in a latency of 30 ms per batch, which is signi\\ufb01cantly faster than ten times the single-image batch inference latency. This is relevant when batching over images captured from different cameras simultaneously. By introducing an improved implementation of convolutional layers in hls4ml, we signi\\ufb01cantly reduce resource consumption, allowing for a fully-on-chip deployment of larger convolutional neural networks. This avoids latency overhead caused by data transfers between off-chip memory and FPGA processing elements, or between multiple devices. Also taking into account the favorable power-ef\\ufb01ciency of FPGAs, we conclude that FPGAs offer highly interesting, low-power alternatives to GPUs for on-vehicle deep learning inference and other computer vision tasks requiring low-power and low-latency. Acknowledgment We acknowledge the Fast Machine Learning collective as an open community of multi-domain experts and collaborators.\"\n        ],\n        \"semantic_type\": \"\",\n        \"description\": \"\"\n      }\n    },\n    {\n      \"column\": \"chunk_char_count\",\n      \"properties\": {\n        \"dtype\": \"number\",\n        \"std\": 590,\n        \"min\": 66,\n        \"max\": 2364,\n        \"num_unique_values\": 34,\n        \"samples\": [\n          1152,\n          992,\n          517\n        ],\n        \"semantic_type\": \"\",\n        \"description\": \"\"\n      }\n    },\n    {\n      \"column\": \"chunk_word_count\",\n      \"properties\": {\n        \"dtype\": \"number\",\n        \"std\": 99,\n        \"min\": 10,\n        \"max\": 425,\n        \"num_unique_values\": 35,\n        \"samples\": [\n          56,\n          134,\n          229\n        ],\n        \"semantic_type\": \"\",\n        \"description\": \"\"\n      }\n    },\n    {\n      \"column\": \"chunk_token_count\",\n      \"properties\": {\n        \"dtype\": \"number\",\n        \"std\": 147.61663313512898,\n        \"min\": 16.5,\n        \"max\": 591.0,\n        \"num_unique_values\": 34,\n        \"samples\": [\n          288.0,\n          248.0,\n          129.25\n        ],\n        \"semantic_type\": \"\",\n        \"description\": \"\"\n      }\n    },\n    {\n      \"column\": \"embedding\",\n      \"properties\": {\n        \"dtype\": \"string\",\n        \"num_unique_values\": 35,\n        \"samples\": [\n          \"[-1.38410693e-02 -4.60955054e-02 -3.22018564e-02  3.14829648e-02\\n -5.36167771e-02 -5.76962307e-02 -2.44085118e-03 -2.64823455e-02\\n -5.56287880e-04 -1.36006847e-02 -2.28371844e-02  3.97218019e-02\\n -6.50844816e-03  6.30740672e-02  1.40205771e-02 -4.69316589e-03\\n  4.46752012e-02  7.31566316e-03 -2.60548294e-02 -6.66951910e-02\\n  6.13027997e-02  1.55195184e-02 -4.25752625e-02  7.81614706e-03\\n -3.37189734e-02  2.73758895e-03 -6.34680614e-02  1.49318222e-02\\n  5.09521849e-02  1.52845690e-02  2.63415258e-02  2.24079695e-02\\n -1.15566133e-02  2.40978003e-02  2.61457035e-06 -6.32778779e-02\\n -3.64307426e-02  3.17521542e-02  7.24352896e-02 -3.26579027e-02\\n -2.27555260e-03  7.09858257e-03 -2.66757160e-02 -3.48306447e-02\\n  4.61809859e-02 -2.75347549e-02 -2.32861694e-02 -2.01110449e-02\\n -4.67833206e-02  1.98740177e-02 -5.95632289e-03  6.66852221e-02\\n  3.51080745e-02  3.94274527e-03  3.03286184e-02  7.22682057e-03\\n  1.85294216e-03 -3.91386710e-02  1.88870239e-03 -2.01514680e-02\\n -1.45996511e-02  1.78687628e-02 -2.15644427e-02  3.91127691e-02\\n  7.22826049e-02  3.40551212e-02  3.68376565e-03  5.80323860e-04\\n -3.92302163e-02  3.35507169e-02 -2.99152900e-02 -1.43148229e-02\\n -1.75864790e-02 -2.72064880e-02 -1.16060004e-01  1.62826162e-02\\n  3.00730416e-03 -1.40209738e-02  5.87847643e-02  1.09720230e-02\\n -3.47371623e-02  2.25245003e-02  6.90930635e-02  5.69968075e-02\\n -9.30649508e-03  7.05549046e-02  1.53897889e-02  1.01945763e-02\\n  2.42398847e-02  1.58079043e-02 -3.31454584e-03  5.04446158e-04\\n  4.95830327e-02  1.83165278e-02  5.45940772e-02 -1.04512265e-02\\n  4.95559303e-03  8.25096946e-03 -2.06375904e-02  4.42925952e-02\\n  7.28940070e-02  3.74802807e-03  2.23079510e-02  3.64597887e-02\\n -4.39653322e-02 -3.70230787e-02 -4.11347821e-02  1.37138516e-02\\n -4.34877649e-02 -5.47356568e-02 -5.08306874e-03  7.89024122e-03\\n  9.11680609e-03 -1.95314083e-03 -5.75464517e-02 -1.65595245e-02\\n  4.39580716e-02 -3.19063989e-03  1.58474017e-02 -3.47769186e-02\\n -2.80997027e-02 -3.30506009e-03 -7.90750701e-03 -6.73271492e-02\\n  2.93593202e-03 -1.09529938e-04  1.47505598e-02 -2.40678005e-02\\n  1.79197472e-02  2.11120513e-03 -6.65894523e-02 -1.82836279e-02\\n  1.36999041e-02  6.40346762e-03 -2.19826326e-02  2.94189546e-02\\n -8.71831924e-03 -2.41784435e-02  8.53949934e-02  1.68274355e-03\\n  5.68606611e-03 -1.67383663e-02 -4.77207778e-03 -1.29108410e-02\\n  4.63540703e-02  3.48802209e-02  1.63939595e-02 -9.69254971e-03\\n  6.38555288e-02 -1.86303742e-02 -9.76920407e-03  4.74817082e-02\\n  1.34980716e-02 -2.13034488e-02 -6.61344733e-03  3.35725700e-03\\n  1.97643787e-02  5.96182644e-02  1.93976928e-02 -1.21373758e-02\\n  2.48643849e-03  3.04449387e-02  1.30743226e-02  3.33384648e-02\\n  6.99523464e-02 -5.26263639e-02 -3.04135159e-02 -8.32764953e-02\\n -1.50272604e-02 -4.12555709e-02  3.23240161e-02  4.06789519e-02\\n  1.84915273e-03 -5.05073108e-02  3.11936848e-02  4.27860282e-02\\n -4.39314591e-03 -4.63537732e-03 -1.48822293e-02 -5.34956865e-02\\n  2.88891955e-03 -7.84085914e-02 -4.05204333e-02  4.15211320e-02\\n -4.24950123e-02 -5.47465347e-02  2.46679038e-02  4.40690219e-02\\n  2.03394406e-02 -1.10826632e-02 -2.02882793e-02  1.00104949e-02\\n  8.94652754e-02 -4.81187291e-02  1.51777118e-02 -1.96196418e-03\\n -2.23784856e-02  1.17917946e-02  1.41518852e-02 -6.52350858e-02\\n  1.78160463e-02 -5.95802628e-02  5.88227920e-02 -5.59113733e-02\\n  4.95304205e-02 -4.77205925e-02 -5.50996065e-02  6.03281446e-02\\n -4.40222956e-02  5.90803195e-03  2.75743604e-02  1.22778174e-02\\n  9.05633047e-02  2.07820497e-02 -2.05460582e-02  8.44789110e-03\\n  5.00034057e-02 -3.69858108e-02  1.23388404e-02  1.27362640e-04\\n  3.09928544e-02 -3.08762845e-02  6.05388209e-02 -2.04069316e-02\\n  1.55486707e-02  6.38128966e-02  4.33013402e-02  7.76233710e-03\\n  9.38763842e-03  1.63593274e-02  6.15334660e-02 -1.58743653e-02\\n  4.27714512e-02 -4.40521538e-02 -2.87858751e-02 -9.18958150e-03\\n  4.39742254e-03  4.15803157e-02 -1.34178707e-02 -2.29825545e-02\\n  6.77823424e-02  2.31545866e-02  7.56444409e-04  4.08224836e-02\\n  5.84859680e-03  5.32506146e-02 -3.89874130e-02 -2.93347109e-02\\n -4.34322730e-02  1.94279905e-02  4.96277735e-02  2.51772590e-02\\n -6.62971381e-03  2.07776632e-02  4.91279773e-02 -2.27969047e-02\\n  3.04034613e-02  1.75280422e-02  6.08424237e-03  2.77555715e-02\\n  4.02762368e-03  2.50704829e-02 -3.01045887e-02  1.68577787e-02\\n -2.18755547e-02 -3.08602750e-02  2.02900041e-02 -7.50091253e-03\\n  4.41387109e-02  9.69169289e-03  8.69436935e-03  3.39226834e-02\\n  3.23769562e-02 -2.10483298e-02  2.77970526e-02  1.38058010e-02\\n  8.41590948e-03  2.00230954e-03 -3.53399999e-02  2.43452080e-02\\n -6.36959225e-02 -6.96382299e-03  2.02840418e-02 -4.02767621e-02\\n -4.84061986e-02 -5.06851450e-03 -6.88633025e-02  9.49658267e-03\\n -2.95379944e-02  1.58555266e-02  9.88566037e-03  2.56013777e-02\\n -3.53269540e-02 -2.16167071e-03 -2.34552827e-02 -2.47176196e-02\\n  2.34885328e-02 -1.03057278e-02 -3.17926146e-03  1.39990775e-02\\n  2.41009817e-02  4.79156449e-02 -9.42198676e-05 -1.33937933e-02\\n -2.86720209e-02  3.43292207e-02  4.42441665e-02 -5.51254302e-03\\n -2.10004020e-02  3.35821509e-02  7.12117739e-03 -4.20419965e-03\\n -1.83078665e-02 -5.63805457e-03 -5.31164035e-02 -1.01054914e-01\\n  1.03214858e-02  8.57730955e-02  5.26594557e-02 -3.89241404e-03\\n  5.06478138e-02  1.80287622e-02  2.46079862e-02 -9.83938668e-03\\n -1.99711118e-02 -2.73605455e-02 -8.15049633e-02  8.70327652e-03\\n  1.78737808e-02  1.59823447e-02  7.04085976e-02  1.71107091e-02\\n  4.04342115e-02 -2.93360054e-02 -6.94894642e-02  8.42152815e-03\\n  3.05859447e-02 -1.74472965e-02  1.72010958e-02 -2.76917424e-02\\n  1.22338980e-02 -2.28665955e-02 -1.86235700e-02 -4.76498768e-04\\n  1.13630714e-02 -4.37420793e-03 -5.19832820e-02 -1.37385167e-02\\n -4.18337435e-02 -3.35906283e-03  2.93741282e-02  2.22499128e-02\\n -2.09620912e-02 -1.10488869e-02 -4.45493758e-02  8.81986767e-02\\n -4.20072824e-02  6.31650537e-03  4.25115973e-03  1.21454895e-02\\n  2.53602248e-02  7.85033628e-02  3.47872898e-02 -2.14430466e-02\\n -8.29541832e-02  4.67981920e-02  2.41953526e-02  2.10421607e-02\\n -9.35274269e-03 -5.01688533e-02  4.09624018e-02 -4.25997414e-02\\n  1.64870284e-02 -3.52229811e-02 -2.80225165e-02 -4.14512120e-02\\n  4.27136384e-03  4.90603298e-02  3.21082473e-02 -4.45395149e-02\\n -4.23068926e-03  2.86526103e-02  1.64737776e-02  2.81865466e-02\\n -2.62129046e-02  4.69130045e-03  2.43143290e-02 -3.23326923e-02\\n -6.18663579e-02  3.40160057e-02  2.37785261e-02  2.88764969e-03\\n -2.92668361e-02 -7.17924759e-02  1.30333323e-02 -2.90943049e-02\\n  3.26396748e-02  1.53876643e-03  1.07401041e-02  2.25557503e-03\\n -5.62781431e-02  1.20429257e-02  3.94307226e-02  1.18236477e-03\\n  2.87926011e-02  2.90144961e-02  2.92756036e-02 -9.48753674e-03\\n -4.11037914e-03 -9.36406571e-03  2.29467377e-02 -3.33455466e-02\\n  3.25238705e-02 -1.19841881e-01 -4.23824191e-02  2.48559695e-02\\n -3.51420082e-02 -1.94709785e-02  4.30667549e-02 -5.52309491e-03\\n -1.14777148e-01 -2.44847164e-02  9.60335806e-02  8.25040229e-03\\n  4.80653252e-03 -4.09438740e-03  3.85695249e-02  8.66774935e-03\\n  6.53093457e-02 -4.96881753e-02 -3.47937420e-02 -5.46298772e-02\\n -1.63763221e-02  3.97519805e-02 -7.57942498e-02 -1.53115029e-02\\n -2.73384955e-02  7.11665824e-02  5.23915561e-03 -7.28224590e-03\\n  4.29212824e-02 -3.47541161e-02 -4.05200720e-02  3.54981748e-03\\n -2.81663425e-02  2.14495994e-02 -2.19514091e-02 -3.62473764e-02\\n -3.24995592e-02 -3.77544835e-02 -1.00759941e-03 -4.91675474e-02\\n -2.46013817e-03 -4.11208533e-03  1.66726094e-02 -1.53967096e-02\\n -2.40476448e-02 -5.54838665e-02 -4.81775440e-02 -1.70829743e-02\\n  7.64188096e-02  8.78789462e-03  6.97883517e-02  4.09519896e-02\\n  4.72494215e-02  3.17292474e-02 -3.02762119e-03 -4.88121528e-03\\n  2.81849150e-02  1.65177416e-02 -1.65478811e-02 -8.09888095e-02\\n  3.74253727e-02 -2.54593864e-02 -2.42416351e-03 -7.08648050e-03\\n  4.98028286e-02  2.05914881e-02 -2.40406878e-02  4.70234081e-02\\n -3.31226289e-02  3.89240645e-02  8.41714628e-03  3.75251211e-02\\n -1.15407640e-02 -3.53405662e-02  1.15647083e-02  1.43297054e-02\\n -2.47821119e-03  1.08115561e-02  2.20704060e-02 -6.62343297e-03\\n  1.27343610e-02 -2.78795827e-02  5.66644706e-02 -2.08238866e-02\\n -5.64980358e-02 -1.52633886e-03  8.14440940e-03 -2.16111969e-02\\n -3.09868553e-03  1.19455121e-02 -5.43401092e-02  1.10017203e-01\\n -6.22302443e-02  2.32332572e-02 -7.10625052e-02  1.59942601e-02\\n -9.60972905e-03  3.27626988e-02 -1.09932110e-01  4.51639704e-02\\n -5.79936840e-02 -2.36959811e-02  2.11754236e-02  4.68421839e-02\\n  9.47354212e-02  8.50492269e-02  7.92326033e-02  2.50817044e-03\\n -3.53150629e-02 -1.13030861e-03  5.51350452e-02  2.59160548e-02\\n  2.70644668e-02 -4.59268168e-02  2.60623707e-03  4.90977755e-03\\n  6.23159111e-02 -4.08030376e-02  5.58380876e-03  1.32920584e-02\\n -1.45420609e-02  2.59492490e-02  3.33277173e-02 -1.95901524e-02\\n -4.74625267e-02 -3.99523191e-02 -4.40525450e-02 -7.53429346e-03\\n -5.78930508e-03  2.65485197e-02  7.00208396e-02  4.73985374e-02\\n -5.37171848e-02 -3.65729891e-02 -1.60634965e-02  7.49742892e-03\\n  1.70529913e-03 -1.03904895e-01  1.68921705e-02 -1.83500499e-02\\n -1.59945730e-02 -4.61055227e-02  1.73530634e-02 -7.18129629e-33\\n -1.50530729e-02  8.50977153e-02 -6.13511615e-02  3.28935795e-02\\n -1.97804477e-02  1.94539111e-02 -2.55590081e-02 -6.17068522e-02\\n  6.14789575e-02  4.73968014e-02  1.15390327e-02 -2.52376068e-02\\n  1.31982882e-02  6.33136230e-03 -3.22225802e-02  3.65558360e-03\\n -3.41054797e-02  3.43613811e-02  5.05628472e-04  2.46167444e-02\\n  7.28583485e-02  2.97177788e-02  4.73341811e-03  4.25610133e-02\\n -1.26179699e-02  5.49304346e-03 -6.52231425e-02  5.29811531e-02\\n -6.23782352e-02  5.97394109e-02 -3.81288826e-02 -5.46921743e-03\\n -2.37374473e-02 -2.06597848e-03  4.00694646e-03  2.42195390e-02\\n -4.73652221e-02  3.06870658e-02 -1.60266813e-02  1.34904310e-02\\n -5.51456958e-02  8.91556777e-03 -4.48985472e-02  1.20441895e-02\\n -1.19468376e-01 -8.70592950e-04 -3.95190297e-03  1.32394657e-02\\n  3.45315747e-02  3.48324068e-02 -4.57615070e-02 -3.97960655e-02\\n -3.80073823e-02  2.63169557e-02 -2.66289599e-02 -4.51931776e-03\\n  2.81722322e-02  1.46504669e-02 -2.46094968e-02  5.31968055e-03\\n -4.16266061e-02 -7.66947679e-03  5.51465601e-02  3.51982154e-02\\n -1.88313518e-02 -3.23276930e-02 -1.61312800e-02 -3.27089219e-04\\n  1.74115431e-02  2.62883957e-02  5.01108654e-02  1.92700680e-02\\n  2.05568504e-02 -1.34017607e-02  5.80952056e-02  2.44681574e-02\\n -5.72004430e-02 -1.35840112e-02  4.29811031e-02  3.43758650e-02\\n -4.70061488e-02  3.87074985e-03 -7.58627150e-03  1.40636060e-02\\n -8.38187896e-03 -3.99864316e-02  6.47583045e-03 -5.49032837e-02\\n -2.42062919e-02  4.29960055e-04 -4.61588018e-02 -4.17968072e-02\\n -7.56842643e-03  1.41584286e-02 -1.55989034e-02  2.57720593e-02\\n  3.64747494e-02 -3.06647718e-02 -5.87963685e-02  1.74838770e-02\\n  5.29336371e-02 -1.53368368e-04 -4.28300872e-02 -1.34514449e-02\\n -8.32677761e-04 -2.86748428e-02 -1.72799616e-03  4.62668203e-02\\n  1.97600713e-03  1.36569440e-02 -2.78291572e-02 -1.50418247e-03\\n  4.38170806e-02  2.83423401e-02  4.11111712e-02 -7.22339153e-02\\n -7.14285532e-03  3.39364484e-02 -2.62901168e-02 -6.10397616e-03\\n  2.84914020e-02  6.05209656e-02 -6.04328793e-03  5.49077876e-02\\n -1.06710028e-02 -2.98083741e-02 -8.07167590e-02 -4.02902402e-02\\n  1.36945359e-02 -2.28857286e-02  5.16438717e-03  1.44519573e-02\\n  3.05717123e-07  4.27836552e-04  1.29935425e-02  1.19934743e-02\\n -1.61611587e-02 -2.52359789e-02 -4.72520851e-03 -2.39748694e-02\\n  1.09079632e-03 -7.94442147e-02 -2.77971104e-02 -8.01862869e-03\\n  2.26984303e-02  2.15588156e-02 -1.22625893e-02  2.37607714e-02\\n -1.97377317e-02  2.05335189e-02  2.83511858e-02  2.23149545e-02\\n -4.32332866e-02 -5.51229306e-02 -4.95952414e-03  5.38908727e-02\\n  4.21321131e-02 -3.52717973e-02 -4.33911495e-02  6.10519573e-03\\n -6.21891161e-03  5.96885867e-02  1.03888838e-02 -3.72219197e-02\\n -6.57164529e-02  5.23160386e-04  1.27301216e-02  1.27733033e-02\\n  3.37149538e-02  2.97296420e-02  1.73286982e-02  7.83940312e-03\\n -1.17744179e-02 -8.56617615e-02  4.10331301e-02 -1.74503531e-02\\n -1.19219413e-02  3.24562266e-02  2.70309858e-02  3.84385977e-03\\n -1.11217853e-02 -5.39511302e-03 -6.71100710e-03  2.08523660e-03\\n  7.56860618e-03 -9.10280365e-03 -4.21376852e-03 -2.73296330e-02\\n -7.42713455e-04 -3.89146917e-02 -2.42693555e-02 -2.25306526e-02\\n -1.06077073e-02 -1.38613407e-03  1.18276319e-02 -8.17423873e-03\\n  1.77188292e-02  4.77574207e-02 -8.04386288e-03 -5.43961674e-02\\n  3.30695994e-34  4.70254049e-02  1.52228596e-02 -5.98952100e-02\\n  2.84467135e-02 -3.77298221e-02 -2.18203315e-03 -3.97703052e-02\\n -8.13187659e-03 -5.17932735e-02 -4.34261793e-03  2.63041984e-02]\",\n          \"[-5.75716682e-02  4.95334230e-02 -1.75185185e-02  3.57321464e-02\\n -1.66555494e-02 -3.88858467e-02  5.70504144e-02 -8.25155433e-03\\n  1.84094943e-02 -1.47623299e-02  1.72739029e-02 -3.39828432e-02\\n -1.93732027e-02  6.75011203e-02  6.96607458e-04  7.13453162e-03\\n  1.52565660e-02  9.96468496e-03 -1.00247804e-02 -4.30993512e-02\\n -2.62842309e-02  1.78643444e-04 -6.74498209e-04  2.54605766e-02\\n -2.66786627e-02  9.27291811e-03 -1.30211953e-02  9.72228590e-03\\n  2.44463291e-02 -1.49618872e-02  1.54371411e-02  1.66774914e-02\\n -2.81805340e-02  1.98976062e-02  2.21005075e-06 -1.36432992e-02\\n  1.16781284e-05 -1.78486225e-04 -1.73374619e-02  3.79822887e-02\\n -3.41337249e-02 -2.50921454e-02 -2.07022317e-02 -3.99643965e-02\\n -8.66229739e-03  4.44485061e-02  9.08575132e-02  3.01011913e-02\\n  1.54969399e-03  5.54944724e-02 -1.49585120e-02  3.70786153e-02\\n  3.73011976e-02 -3.25192027e-02  4.97213984e-03 -8.01600441e-02\\n  1.16199758e-02  1.57399732e-03 -6.32332778e-03 -4.94671501e-02\\n -4.58032601e-02  6.03061169e-02  6.04039095e-02 -2.52119265e-02\\n  8.71034116e-02 -5.39251463e-03 -5.59890568e-02 -2.20555384e-02\\n  5.94870932e-03  3.32655013e-02 -1.09848380e-01 -1.29265394e-02\\n  2.84378882e-02 -3.06123141e-02 -7.42609128e-02 -1.54590774e-02\\n -1.46787744e-02 -4.14147973e-02  2.08257902e-02  1.71416476e-02\\n -4.31748703e-02  2.27906834e-02  3.94847095e-02  3.13523225e-02\\n -1.81888957e-02  2.80070063e-02  1.01018986e-02  1.13549866e-02\\n  1.55841056e-02  1.81213971e-02 -2.69572530e-02 -2.65848171e-02\\n -9.93534178e-03 -4.42814752e-02 -3.54188792e-02 -5.17274812e-02\\n -1.49654206e-02  1.43108089e-02 -4.74366546e-02 -2.65087988e-02\\n  2.38742549e-02  6.59791157e-02  1.08746858e-03  1.00901863e-02\\n -5.33490293e-02  1.62439086e-02 -3.90300713e-02  1.27759418e-02\\n -4.21855450e-02  8.26233253e-03  1.35040414e-02  9.15834308e-03\\n -2.42548157e-02 -3.19508673e-03  2.44810572e-03 -1.64852142e-02\\n -6.98732808e-02  4.08168696e-02 -4.38095778e-02 -4.52900380e-02\\n -1.54102445e-02  6.57198951e-03 -4.23015654e-02  2.46717338e-03\\n -3.62463668e-02  2.10574362e-03 -1.10128811e-02 -8.99604801e-03\\n -7.81102059e-03  2.50758277e-03 -1.51361544e-02  1.16642695e-02\\n  2.34679151e-02 -3.41352783e-02  3.36111337e-02  6.51383586e-03\\n -4.38607577e-03 -2.52312496e-02 -2.61181872e-03  2.64722984e-02\\n -6.81219343e-03 -4.11103666e-02  5.32725453e-03  5.63349575e-03\\n  1.94358695e-02 -3.04573700e-02  2.31702048e-02  3.59196030e-03\\n -2.30238475e-02 -4.59787995e-02 -1.33287758e-02  1.18312407e-02\\n  3.59411538e-02 -3.50522958e-02  6.11093566e-02 -4.01842855e-02\\n -6.78954460e-03  1.19981892e-01 -2.03330885e-03  7.31218830e-02\\n -2.77146697e-03  4.91611054e-03 -5.70576079e-03 -3.29002598e-03\\n  4.72472943e-02  3.64390644e-03 -8.03636238e-02 -6.70245141e-02\\n -1.41442101e-02  3.60108018e-02  9.12487060e-02 -1.90781392e-02\\n -7.61280060e-02 -4.45698053e-02 -2.39607170e-02  3.09318565e-02\\n  6.08324371e-02 -1.78678576e-02  1.13282781e-02 -1.46779884e-02\\n -8.20827577e-03 -3.22379321e-02 -4.99857590e-02  9.91902947e-02\\n  2.44868379e-02 -1.47068442e-03  5.71063831e-02  6.55379891e-02\\n -3.27542312e-02 -4.53043282e-02  1.97465420e-02  3.57384235e-02\\n -6.29656762e-02  2.62808856e-02  8.15073848e-02 -3.32964025e-02\\n  7.80149410e-03  3.04149762e-02 -5.02781244e-03 -4.76409076e-03\\n -2.49346346e-02 -9.22902077e-02 -1.94959622e-02  1.99919641e-02\\n  4.67077456e-02 -4.12915647e-03 -7.14838738e-04  4.21096310e-02\\n -3.81197818e-02 -3.76504846e-02 -4.46534436e-03  4.03227173e-02\\n  4.61620912e-02 -1.66649129e-02 -4.62174974e-02 -1.99011751e-02\\n  1.60399619e-02 -5.67997992e-03 -2.02317592e-02  3.15231346e-02\\n  5.71878720e-03  1.07801417e-02  9.31081083e-03  5.44985151e-03\\n  4.02678102e-02 -2.61740144e-02  6.67833015e-02  1.52273998e-02\\n  1.63491219e-02  8.06986354e-03 -8.13193340e-03 -2.29169596e-02\\n  8.42592306e-03 -3.86244915e-02  7.22574815e-03 -4.20887247e-02\\n  1.23558370e-02 -4.69110114e-03  1.14009883e-02  2.73935642e-04\\n  3.99907567e-02  8.53408594e-03 -2.74246261e-02 -5.08631915e-02\\n  6.82915300e-02  6.00518249e-02 -1.50169218e-02 -4.63015810e-02\\n  5.78857107e-05  4.19464544e-04  1.97346043e-02  1.86637230e-02\\n  9.13599040e-03  6.23070309e-03 -3.00142588e-03  2.19395105e-03\\n  2.43720016e-03  1.23598976e-02 -6.96875751e-02 -5.36684692e-03\\n  9.72561166e-03  8.29412322e-03  2.80810371e-02  3.77772152e-02\\n -3.70310037e-03  4.72424813e-02  4.24745306e-03 -4.86456864e-02\\n  2.63349656e-02 -3.86205688e-02 -1.26722492e-02 -2.02703830e-02\\n  2.64198873e-02  2.29706289e-03  1.96906179e-03 -2.41680653e-03\\n  2.63578189e-03 -4.44943376e-04 -1.09993825e-02 -5.94310574e-02\\n -2.80056242e-03 -1.73786338e-02 -6.30730316e-02 -8.05734619e-02\\n  1.93844475e-02  5.63006140e-02 -3.45497616e-02 -5.05626313e-02\\n  4.89080884e-02 -5.46029210e-02  8.81525408e-03  2.61798501e-02\\n -4.49415669e-03 -4.09313142e-02 -2.10445076e-02 -3.02212201e-02\\n -6.39746636e-02  1.19354233e-01 -1.91593766e-02  2.61586066e-02\\n  5.15066758e-02  2.97134127e-02  1.25166383e-02  5.16925827e-02\\n -4.63399030e-02 -1.28843309e-02  1.73484981e-02  2.28460087e-03\\n -1.53949885e-02  7.73298666e-02 -2.65988940e-03 -2.58057918e-02\\n -4.17386778e-02  7.74743455e-03  4.76563256e-03 -1.27514265e-02\\n -6.98389634e-02  1.21516304e-03 -1.60424076e-02  5.88268861e-02\\n  9.83521808e-03 -6.12036278e-03  1.12356404e-02  2.82602403e-02\\n  2.12472696e-02  3.75369610e-03 -1.41032189e-02  3.34864995e-03\\n -1.48405163e-02  4.59032767e-02  5.86316437e-02 -3.95315737e-02\\n  4.05867770e-02 -5.81045519e-04  2.19083447e-02  1.56050371e-02\\n -2.31713080e-03 -1.30996406e-02  4.27067745e-04  4.01549339e-02\\n -2.81412597e-03 -1.60204805e-03 -9.94875561e-03  5.35671487e-02\\n -1.76103469e-02 -8.09934514e-04 -1.03416992e-02 -1.76291820e-02\\n  2.03062943e-03  3.25560085e-02 -7.35741705e-02  8.84684026e-02\\n -5.04223406e-02 -1.58727039e-02 -2.91846395e-02  2.73605771e-02\\n  3.74086760e-02 -2.04779115e-02  1.79075506e-02  5.08169569e-02\\n  7.17801899e-02 -3.60939093e-02  5.18979579e-02  4.50964831e-02\\n -5.16346805e-02  4.16626893e-02 -1.06739383e-02  2.25597564e-02\\n -4.33497019e-02 -1.96536514e-03  2.24596485e-02 -4.67717387e-02\\n  1.07615320e-02  2.65026614e-02 -1.64101273e-02 -3.17012221e-02\\n -1.27155916e-03  3.45887169e-02  2.26528533e-02 -6.15642406e-03\\n  3.61182541e-02 -7.06304908e-02  5.48108248e-03 -4.53153066e-03\\n  9.59239528e-03  8.00152123e-03  3.56538966e-02  3.21143493e-02\\n  1.61825269e-02  5.39780706e-02  1.96187813e-02 -2.14179652e-03\\n  1.32552432e-02  9.51704336e-04 -4.08794880e-02 -8.58276114e-02\\n  2.52481792e-02  4.52059694e-03  2.75276173e-02 -3.27893533e-02\\n  1.60461676e-03  5.55236973e-02 -5.37374765e-02 -1.18015055e-03\\n -9.56191123e-03  2.23403573e-02  9.97788738e-03 -1.64445080e-02\\n  3.61182680e-03 -4.43856828e-02 -2.04547867e-02 -4.22318801e-02\\n -6.07441273e-03 -1.15289260e-02  1.87809449e-02  2.10921206e-02\\n -4.29487079e-02 -6.91421516e-03  6.69232709e-03  3.85115705e-02\\n -1.35390703e-02 -6.31982386e-02  4.26568687e-02 -4.40201536e-02\\n -6.16603484e-03 -1.74477492e-02 -2.23551150e-02 -5.71417548e-02\\n  1.59551688e-02 -1.04486179e-02 -1.70677137e-02 -3.38263996e-02\\n -1.53410025e-02 -3.10843647e-03 -8.85903686e-02  5.35620283e-03\\n -8.91894195e-03 -3.99165675e-02 -9.29625332e-02 -3.04030832e-02\\n  5.75126819e-02 -2.31690202e-02  8.15184191e-02 -3.77199128e-02\\n -7.07119191e-03 -1.80301387e-02  3.30180526e-02 -3.83264683e-02\\n  5.96416853e-02  3.73022631e-02  8.16167984e-03 -1.88800134e-02\\n -3.85422702e-03  5.37143927e-03 -2.40627043e-02 -3.89825888e-02\\n -1.05629852e-02 -6.62031621e-02 -6.24577962e-02 -1.13444142e-02\\n  4.50660884e-02  1.77512132e-02  5.55683672e-02 -5.03934361e-03\\n -6.38734177e-02  5.62388264e-02  2.47446988e-02 -2.82874443e-02\\n  3.47172841e-02  3.76123749e-02 -5.86044751e-02 -8.73540714e-02\\n  2.21916251e-02 -3.67172621e-02  6.60940856e-02  1.52011821e-03\\n  1.69719197e-02 -6.06112108e-02 -2.27151103e-02 -4.44486663e-02\\n -4.84340861e-02  8.59529004e-02  3.44752334e-03  4.08127829e-02\\n -4.11221609e-02  9.92145203e-03  4.29019034e-02  4.88900095e-02\\n -5.27807660e-02  5.70012350e-03  2.78846230e-02 -2.22600088e-03\\n -7.98070745e-04 -3.91741432e-02  5.83108142e-02 -1.16138682e-02\\n  2.67506074e-02 -4.21289206e-02 -3.19971405e-02  3.87946703e-02\\n  2.99081090e-03 -7.68902013e-03  1.30951479e-02  1.68930553e-02\\n  4.87780757e-02  3.32473107e-02 -2.41147764e-02  2.44630780e-03\\n -6.30621193e-03  1.77534185e-02 -6.39839396e-02  5.05636074e-03\\n -4.07015756e-02 -6.90498576e-02 -7.02026039e-02 -3.36020105e-02\\n  6.19614907e-02 -1.18141873e-02  3.05484589e-02  6.65274449e-03\\n -2.50640288e-02  5.54247480e-03  1.28422799e-02  4.60093692e-02\\n  4.95589562e-02  3.11537609e-02  2.69439854e-02  1.40465554e-02\\n -2.65042130e-02  7.42021296e-03  4.89413291e-02  4.79415711e-03\\n -2.47168858e-02  1.99942347e-02 -1.88126171e-03 -1.09314872e-02\\n -2.32622046e-02  7.66003504e-02  9.43633448e-03 -7.10792979e-03\\n -5.66471834e-03 -5.62403537e-03  4.95352522e-02  8.11232161e-03\\n -1.85943544e-02 -4.28635180e-02  1.36266202e-02  8.69223103e-03\\n  4.85231215e-03 -4.78061326e-02 -1.69536117e-02 -6.36829212e-02\\n  1.34961046e-02  1.62619594e-02  2.38747690e-02 -6.38137239e-33\\n -1.36307012e-02  9.72431246e-03  6.82674255e-03 -7.35139241e-03\\n -1.08659444e-02  3.22139859e-02  1.11747617e-02 -2.39784922e-02\\n -1.29451668e-02  1.49589609e-02  2.89139367e-04  8.80814437e-03\\n  3.00790053e-02 -1.98854674e-02  9.75116715e-03  3.04625891e-02\\n  7.98010156e-02 -1.76952630e-02  7.40322005e-03  3.83047424e-02\\n -2.25320347e-02  3.04292305e-03  2.75082309e-02  1.83611661e-02\\n -5.10496497e-02  5.49643002e-02 -2.99298894e-02 -6.99415989e-03\\n  1.91048179e-02  4.34989035e-02 -4.12848443e-02  2.01926921e-02\\n -4.76506166e-02 -6.64191786e-03  7.20163062e-03 -3.69789079e-02\\n -5.66058382e-02 -2.42689718e-02 -1.29706450e-02 -6.48761773e-03\\n -2.50547044e-02 -1.98807381e-02  1.23115396e-02  4.56986539e-02\\n -5.23833707e-02 -3.11576985e-02  6.01013284e-03  2.48365141e-02\\n -4.61230054e-02 -7.01086149e-02  9.06275492e-03  5.17327748e-02\\n -1.91623233e-02  1.70711372e-02  7.87387267e-02 -1.78942271e-02\\n -3.53354551e-02 -2.94880215e-02 -5.42037450e-02 -2.99784597e-02\\n  2.53677964e-02  1.23852286e-02  4.71462533e-02  5.28991334e-02\\n  4.54172045e-02  5.34699000e-02  4.22706455e-02 -1.02775246e-02\\n  2.69788746e-02  1.00162216e-02  4.36448269e-02  5.62659465e-02\\n  2.89022215e-02 -1.06295086e-02  8.64569694e-02 -4.12609726e-02\\n -4.63470705e-02  7.93899584e-04 -2.11413465e-02  1.98454242e-02\\n  6.77707233e-03  5.44372275e-02 -2.06225850e-02 -1.55005809e-02\\n -2.62684487e-02 -2.13259459e-02 -3.00401077e-02 -1.59540586e-02\\n  7.35750562e-03  4.34464104e-02 -3.88678722e-02  6.63680024e-03\\n  3.15750502e-02 -1.69465393e-02  8.71863142e-02 -1.52613390e-02\\n -1.59710404e-02  3.23828869e-02  5.87724186e-02  9.08884592e-03\\n -2.45841984e-02 -1.45629253e-02 -7.17201410e-03 -1.99283063e-02\\n -1.49131753e-03 -4.50732652e-03 -1.71798859e-02  6.17028512e-02\\n -1.46524711e-02 -1.40636815e-02 -3.75044197e-02  2.01276913e-02\\n  4.93611433e-02  5.79230562e-02  1.60642862e-02  8.15755501e-03\\n  6.67552091e-03  7.52551258e-02 -6.05071113e-02  1.46316076e-02\\n -8.57051928e-03  3.34465727e-02  4.91540469e-02  6.12262152e-02\\n -5.34312092e-02 -3.03371716e-03 -4.01012115e-02 -3.58650014e-02\\n  2.18644422e-02 -1.14184514e-01 -1.83837302e-02  4.87794131e-02\\n  2.83739752e-07 -1.37539301e-02 -1.88800860e-02 -5.49100302e-02\\n -2.36638263e-02 -5.36661781e-03 -8.46315641e-03  3.73495370e-02\\n -4.60294038e-02  4.20938199e-03  7.49173900e-03  1.86448656e-02\\n  3.44446823e-02  3.50322723e-02 -7.38524459e-03 -1.17259966e-02\\n  3.62721570e-02  1.50549663e-02  4.15792838e-02 -3.26641835e-02\\n -7.59705752e-02 -4.41951789e-02  6.61871284e-02  1.41602615e-02\\n -1.55055160e-02  7.95357823e-02 -3.85143235e-02  9.61323618e-04\\n  3.86219695e-02  3.83055881e-02  6.78994209e-02  8.03162456e-02\\n -1.50629133e-02  9.85824596e-03  3.24249789e-02  1.69173125e-02\\n  3.62970755e-02 -1.31085189e-02  3.29252370e-02 -2.52219792e-02\\n -7.72906914e-02  8.56630877e-03  3.72621566e-02  2.22893804e-02\\n -5.46494685e-02  5.97084761e-02  9.73945949e-03  6.70672487e-03\\n -9.07882303e-02  6.40410464e-03 -5.08126318e-02  3.36143710e-02\\n -1.01533877e-02  8.11373442e-03  3.28623168e-02  1.88888926e-02\\n -3.32345217e-02  1.59401968e-02 -4.94522303e-02  1.02954376e-02\\n  3.75317931e-02 -9.01888032e-03 -3.83462124e-02 -6.64028302e-02\\n -2.67367475e-02  5.82179129e-02  5.24488613e-02 -4.16087583e-02\\n  2.41940339e-34  6.65245056e-02 -2.83324737e-02  9.38807707e-03\\n  5.98444678e-02  1.58018004e-02 -6.10353053e-03 -5.99875078e-02\\n  7.41486577e-03 -9.33913887e-03 -6.58628866e-02 -2.71459739e-03]\",\n          \"[-3.16746645e-02  3.61035243e-02 -5.64813754e-03  6.76951408e-02\\n -2.43329792e-03  7.93138985e-03  5.99785559e-02 -1.92859769e-02\\n -6.26338795e-02  7.03719817e-03  2.93616783e-02  1.65558774e-02\\n  2.81488965e-03  4.75769490e-03  2.92897737e-03 -2.04686355e-02\\n  1.65884453e-03  3.03444751e-02 -2.53163353e-02 -2.93697529e-02\\n  4.47337814e-02 -1.49275437e-02  2.27272473e-02  2.52604168e-02\\n -3.25208642e-02  1.17384214e-02 -4.98154536e-02  2.31069699e-02\\n  3.46076712e-02 -1.78039558e-02 -8.21412355e-03  1.22018773e-02\\n  4.30541411e-02  8.73358175e-03  2.15362661e-06  1.23219714e-02\\n -2.34571919e-02 -2.28275489e-02 -7.98378233e-03  3.85243702e-03\\n -1.49913307e-03 -1.12337582e-02 -4.76928614e-02 -2.92181764e-02\\n  2.56175492e-02 -6.61158562e-02  9.49366614e-02 -5.73057681e-03\\n  1.62560947e-03  2.06233375e-02  5.34827821e-03 -3.20781954e-02\\n  2.88451128e-02 -1.51191102e-02 -3.30076739e-02 -1.91666111e-02\\n -5.41909831e-03 -3.43508869e-02 -5.04569523e-02 -4.06473689e-02\\n  2.05934048e-03  2.62677111e-02  6.45170808e-02  1.97182689e-02\\n  1.11052804e-02  1.67547967e-02 -3.73367332e-02 -1.84143689e-02\\n  9.67449858e-04  2.56503653e-02 -6.75478801e-02 -1.89861245e-02\\n  2.70446442e-04 -9.75602213e-03 -5.00142872e-02 -7.16825724e-02\\n -6.39195275e-03  3.67797874e-02  4.44722245e-04  3.31041496e-03\\n -5.34689277e-02 -1.78693011e-02  4.16929051e-02  3.52987312e-02\\n  4.59932163e-03 -3.15848179e-03  4.15787995e-02  7.22351372e-02\\n -1.81177333e-02  2.34499127e-02 -4.86865044e-02 -5.98050952e-02\\n -1.08730458e-02  2.31971834e-02 -2.06468124e-02 -1.09382998e-02\\n -6.54302090e-02 -3.90097126e-02 -4.19226810e-02 -2.06105486e-02\\n -4.85185646e-02  1.36278905e-02  2.73071248e-02  3.56993861e-02\\n -2.80852951e-02 -8.55840743e-03 -3.18429396e-02  4.02477980e-02\\n -3.42256092e-02 -4.77935672e-02  1.68956723e-02 -1.13155944e-02\\n -5.69012538e-02  5.41341119e-03 -2.86825672e-02  2.10163109e-02\\n -4.14940715e-02 -3.83577216e-03 -2.05718167e-02 -2.17066333e-02\\n -6.70757070e-02  4.03024405e-02 -2.55645532e-02 -3.26916091e-02\\n  6.19077543e-03 -1.51784555e-03 -1.22181315e-03  3.47758806e-03\\n  2.24173684e-02 -6.14128672e-02  2.28672531e-02  3.48707242e-03\\n  1.95441078e-02  1.98473204e-02  6.05480419e-03  1.07514774e-02\\n -4.98289913e-02 -6.68070018e-02 -6.83642328e-02  3.94121557e-02\\n  6.36559576e-02 -2.61165667e-02  5.16009778e-02 -5.27653173e-02\\n  1.78192724e-02 -5.82608208e-03  3.18697095e-02  1.27865089e-04\\n  3.75285815e-03  2.48754639e-02 -1.76434573e-02  5.69977574e-02\\n -6.89609125e-02 -3.12244892e-02  3.56211215e-02 -7.68751372e-03\\n  4.49266508e-02  9.49850753e-02  5.97571880e-02  4.36045863e-02\\n  1.45970294e-02  2.15809681e-02 -4.29957286e-02  2.55598240e-02\\n  1.29946449e-03 -6.25328301e-03 -6.62354231e-02 -2.29815505e-02\\n -3.01079769e-02  4.04155068e-02  2.83989329e-02 -1.46099953e-02\\n -4.07727025e-02 -3.81505527e-02  1.54687529e-02  2.24641487e-02\\n -3.01146079e-02  4.51189689e-02 -3.09947412e-02  2.45260987e-02\\n -4.16310690e-03 -8.44279677e-02 -2.70167384e-02  5.90071641e-02\\n  3.43906470e-02 -3.06145623e-02  1.39058987e-02  1.71897952e-02\\n -1.95062589e-02 -1.25115421e-02 -1.13685010e-02 -4.14335690e-02\\n  3.71918045e-02 -9.34039243e-03  3.09888180e-02 -4.28426191e-02\\n  2.56936867e-02  2.37612743e-02 -3.72928171e-03  1.79179870e-02\\n -5.74689955e-02 -6.16443753e-02  4.70675528e-02  2.08514649e-02\\n  1.00419228e-03 -1.44619942e-02 -1.99198984e-02  4.36162129e-02\\n  1.24005452e-02 -3.05563454e-02 -2.03936622e-02 -6.02313271e-03\\n  5.41093983e-02 -5.72129115e-02 -9.50805470e-03  2.01492384e-02\\n -1.81069213e-03  1.21593233e-02  3.08283269e-02  5.35118347e-03\\n  1.57063883e-02  6.00888692e-02  2.94326339e-02  1.29250884e-02\\n  5.56495562e-02  4.39563207e-02  2.57587750e-02  4.19635288e-02\\n -1.69137567e-02 -4.42940146e-02 -1.13006718e-02 -4.11034748e-02\\n  3.71334516e-02 -1.89474113e-02  5.16503863e-02 -4.16901112e-02\\n -4.64470983e-02  3.54141602e-03  1.33084059e-02 -5.05993590e-02\\n  1.59035362e-02  6.23197332e-02 -8.18567649e-02 -8.15114193e-03\\n -4.76432778e-03 -1.57912020e-02  1.84115525e-02 -1.84766704e-03\\n -1.10557815e-02  3.66139375e-02  2.73377844e-03  2.95194704e-03\\n  7.39491684e-03 -1.52493380e-02  8.31364468e-03  3.00173517e-02\\n  1.08236382e-02  2.98125446e-02  1.23873912e-02 -6.11370383e-03\\n -1.22917546e-02 -5.96908852e-03 -3.35102826e-02 -6.39618328e-03\\n -1.42085813e-02  2.84412373e-02  7.45144635e-02 -5.60570918e-02\\n  3.92864225e-03 -6.51184693e-02 -2.46089529e-02  1.43751791e-02\\n  2.32991595e-02  4.63746302e-02 -3.35901566e-02  1.29792169e-02\\n  8.71372074e-02 -7.59621710e-03  9.72660352e-03  7.96370301e-03\\n -4.07744525e-03 -7.36734364e-03  1.43551240e-02 -5.97581603e-02\\n  3.97756398e-02  2.60108095e-02 -1.70690604e-04 -6.23490401e-02\\n -4.59900033e-03 -8.07224661e-02 -4.37714010e-02  1.61878113e-02\\n -1.86588466e-02 -3.94625515e-02  3.81548726e-03  1.25734424e-02\\n -3.75268087e-02  1.10110238e-01  3.37748490e-02 -3.32692708e-03\\n  3.78371775e-02  4.46357206e-02  3.40076461e-02  6.79850718e-03\\n  1.77742708e-02  1.62886530e-02 -7.67363142e-03  4.24502157e-02\\n -9.98785347e-02  3.39874141e-02 -5.50876260e-02  1.35049913e-02\\n -3.98010435e-03 -2.63269041e-02  3.30733098e-02 -6.03447221e-02\\n -5.12328744e-02  6.69900514e-03  5.90635603e-03  1.56578943e-02\\n -4.09697071e-02 -5.17048836e-02 -7.69575033e-03 -5.39406808e-03\\n -1.90479290e-02  3.47716101e-02 -1.02834022e-02  1.14692301e-02\\n  3.48722227e-02  6.81458637e-02  4.38925326e-02 -5.53757325e-03\\n  2.98870783e-02  9.34812147e-03 -1.92715731e-02 -5.38648777e-02\\n  8.65929388e-03  5.64715378e-02 -4.20581736e-02  1.87192820e-02\\n -6.82432353e-02 -3.96694802e-02  9.85455420e-03  1.15675731e-02\\n -4.96272556e-02  3.78336832e-02  1.42227747e-02 -2.02298127e-02\\n  4.00133543e-02  4.49804366e-02 -1.81728266e-02  5.04780523e-02\\n  1.64751541e-02 -1.09643098e-02 -1.54574420e-02  2.36047618e-02\\n  3.98497842e-02  5.19343689e-02  3.75977419e-02  2.26069298e-02\\n -3.13175935e-03  6.08279333e-02  2.46510636e-02 -2.42182910e-02\\n -1.28452018e-01  1.35556161e-02  6.55203778e-03  1.87418368e-02\\n -3.21488157e-02  2.41558645e-02  1.52879935e-02 -6.57354742e-02\\n -1.35331657e-02 -1.19685347e-03 -1.53256366e-02 -1.56167587e-02\\n  5.66355092e-03  7.12433234e-02  6.26006396e-03 -3.34743294e-03\\n -1.55424094e-02 -6.26417920e-02 -6.74238503e-02  3.91137488e-02\\n -2.96404045e-02  3.18640913e-03  1.46671459e-02 -1.16221132e-02\\n -1.10561689e-02  8.05947557e-02 -2.05092579e-02 -1.73177272e-02\\n -2.24099532e-02 -1.54217193e-02 -1.31263165e-03 -2.06328817e-02\\n -1.21467113e-02 -3.06138396e-02  1.76454373e-02 -2.49452218e-02\\n -1.85080413e-02 -1.06569454e-02  9.30379257e-02  1.35255232e-02\\n  9.45573766e-03 -5.22037409e-02  4.45102900e-03 -1.26663586e-02\\n  2.84529990e-03 -6.40696287e-02  6.42430224e-03  1.26529681e-02\\n -2.40246616e-02 -8.55875108e-03  3.05212624e-02 -1.97017416e-02\\n -4.37256582e-02  8.18422530e-03  1.45410756e-02  3.70019227e-02\\n -4.48306091e-03 -2.56600920e-02  1.21717285e-02  1.83149672e-03\\n  4.53542136e-02 -1.98948253e-02 -2.66326629e-02 -3.44078280e-02\\n -3.92837226e-02 -3.13082226e-02  2.57424004e-02 -4.90424037e-02\\n -9.59934201e-03 -2.47269087e-02 -7.66595034e-03  4.24237102e-02\\n -4.04466875e-02  8.28601513e-03 -6.07855953e-02  3.11880112e-02\\n -1.97013244e-02  1.74703486e-02  4.16600183e-02 -3.89351211e-02\\n -5.84902540e-02  6.27648011e-02  5.80507889e-02 -1.13679238e-01\\n  2.52884738e-02 -1.96748395e-02  5.31916367e-03 -4.15063016e-02\\n  2.06999928e-02  3.22196782e-02 -2.26651542e-02 -2.41246782e-02\\n  7.57783279e-02 -1.06750824e-01 -1.37048280e-02  1.50517393e-02\\n  8.89685994e-04  5.80889813e-04  3.21493186e-02  2.59409640e-02\\n -2.09744386e-02  3.73607725e-02  3.49422768e-02  1.16999680e-02\\n  2.01365892e-02  5.83995460e-03 -4.47696149e-02 -1.65007338e-02\\n  5.23326769e-02 -1.28071392e-02  4.27743718e-02  1.07490588e-02\\n -4.56076451e-02 -2.11919267e-02 -4.76235300e-02 -2.45459285e-02\\n -1.58191216e-03  9.51817930e-02  2.59720571e-02  4.99429889e-02\\n -6.49667764e-03  6.59544021e-02  6.71816021e-02  2.76763663e-02\\n -5.06020114e-02 -9.63848736e-03 -1.00446129e-02  4.60154552e-04\\n  6.51707640e-03  1.93203464e-02  1.38799921e-01 -3.31417732e-02\\n  5.09953173e-03 -1.28529472e-02 -2.24176422e-02  1.52827045e-02\\n -5.30774444e-02 -5.49711473e-03  2.93215527e-03 -2.30836608e-02\\n  1.87622141e-02  1.84326321e-02 -4.75265607e-02  2.20878981e-02\\n  3.25109921e-02  5.08271996e-03 -4.73281555e-02  5.21334335e-02\\n  7.61335017e-03 -3.19396928e-02 -6.93711564e-02  2.57256776e-02\\n  5.15056700e-02  7.27693667e-04 -5.13713481e-03 -3.14112343e-02\\n -9.65155268e-05 -4.59693372e-02  2.63827052e-02  8.42548069e-03\\n -1.59064792e-02  6.78458288e-02  1.10950356e-03  1.99257974e-02\\n  5.13909943e-03 -1.12409480e-02  9.17013548e-03 -5.31213591e-03\\n  4.55361903e-02  1.93615258e-02 -1.86165571e-02 -4.72357236e-02\\n  4.37691854e-03  1.24608027e-02  2.30293125e-02 -2.26009190e-02\\n  4.99893911e-02  1.09012211e-02 -1.59858093e-02  7.16206734e-04\\n -1.26386508e-02 -1.49600850e-02  1.94098875e-02  4.03419770e-02\\n  1.25403441e-02 -1.14489481e-01  3.67074087e-02  5.70242992e-04\\n -2.88387164e-02  3.78838815e-02  3.72974351e-02 -6.21002131e-33\\n -1.37619339e-02  2.89412457e-02  5.41263372e-02 -2.72149104e-03\\n -8.69605318e-03  2.39095725e-02  2.63885446e-02 -1.71462130e-02\\n -3.34087983e-02  2.06786068e-03 -3.55573371e-02  1.26592517e-02\\n  3.96798458e-03  2.34523974e-02 -4.68763756e-03 -2.84030158e-02\\n  5.13238944e-02 -2.60382164e-02 -1.63293239e-02  2.25878973e-02\\n  1.43691273e-02  7.49988994e-03  4.49874215e-02  3.79288457e-02\\n -9.73533690e-02  5.32927178e-03  9.74865258e-03 -3.82607169e-02\\n -2.48585455e-02  2.30358751e-03 -2.83290856e-02  1.54509610e-02\\n  2.45087259e-02 -7.80091658e-02 -1.13475276e-02  3.43550295e-02\\n -1.14107467e-01 -7.08650891e-03 -1.14016579e-02  6.68505952e-02\\n -4.38594818e-02  1.71223117e-04  6.23384863e-02  4.76352796e-02\\n -6.54618430e-04 -2.72011999e-02 -2.35182587e-02 -1.56553388e-02\\n -3.12306490e-02 -4.07394879e-02  2.65487675e-02  2.49035712e-02\\n  1.74305122e-02  5.25885541e-03  8.46434906e-02 -3.23898382e-02\\n -5.17333150e-02  2.15372071e-02 -5.81636578e-02 -1.56580098e-02\\n -6.61056908e-03  3.89969535e-02  5.73528633e-02  4.37972136e-02\\n  3.24502066e-02  4.21965234e-02  3.35489511e-02 -3.97098772e-02\\n -1.13038933e-02  7.42306262e-02 -6.79414347e-03  3.99883389e-02\\n  1.90526759e-03  2.77155694e-02  9.87365767e-02 -8.88851471e-03\\n -8.95759538e-02 -3.47619951e-02 -1.93603691e-02  3.11617274e-02\\n  6.12067338e-03  7.90808164e-03 -1.41521795e-02  6.09852420e-03\\n  2.02010255e-02  3.05661029e-04  1.98238920e-02 -1.32248383e-02\\n  8.13502632e-03  4.60180305e-02 -5.20170815e-02  2.51165703e-02\\n  2.77545210e-03 -3.57689559e-02  5.31931147e-02  2.85512358e-02\\n  4.13421243e-02  1.75141711e-02  5.83492331e-02  3.23072970e-02\\n -7.96818212e-02 -5.31088561e-03 -2.81711426e-02 -8.46965052e-03\\n -1.44967549e-02 -1.74123161e-02 -8.33824649e-03  4.70952764e-02\\n -2.49125566e-02  1.01361969e-04 -4.71436717e-02  1.85817368e-02\\n -7.10025057e-03  3.39241512e-02  4.10902388e-02 -3.30000208e-03\\n  2.82850079e-02  2.22758111e-03 -7.19814748e-02 -1.49258217e-02\\n -3.36849364e-03  2.12581665e-03 -1.71837173e-02  1.26391044e-02\\n -1.91223826e-02 -1.59826837e-02 -6.03155047e-02  6.79021981e-03\\n  2.81202570e-02 -6.55817240e-02 -8.65310524e-03  3.53386998e-02\\n  2.97365716e-07 -6.27783239e-02 -9.65350214e-03 -6.20612700e-04\\n -3.14228758e-02  1.51277045e-02  2.30126474e-02  3.93226221e-02\\n  3.52786332e-02 -2.59918422e-02 -4.86303726e-03  1.81318466e-02\\n  4.84020486e-02  8.40129238e-03 -9.56198014e-03 -2.21983111e-03\\n  3.15968841e-02  1.11920480e-02  2.19263304e-02 -1.29026091e-02\\n  2.68466468e-03  1.37388306e-02  1.14613488e-01  2.36437395e-02\\n -3.12437005e-02  6.58700019e-02  2.43909787e-02 -3.29759270e-02\\n  2.29644831e-02  3.04527879e-02 -1.66343283e-02  6.38207840e-03\\n  6.31437153e-02 -3.66985761e-02  6.56475797e-02  5.13081588e-02\\n -3.39419171e-02  1.00854726e-03  6.34812936e-02 -1.95487197e-02\\n -3.85944918e-02  1.64904669e-02 -4.13667150e-02 -2.81776823e-02\\n -2.90046688e-02  4.52383235e-02  1.97405424e-02 -1.91350735e-03\\n -5.88023923e-02 -8.23135301e-03  7.36641092e-03  9.64698195e-03\\n  5.55581152e-02  4.87777144e-02 -1.98660865e-02 -6.46073790e-03\\n -7.09221214e-02 -1.69412363e-02 -5.46945482e-02  2.47643981e-02\\n -1.77909117e-02 -2.25497261e-02 -4.58168685e-02 -8.67442712e-02\\n  5.12560690e-03  5.23791611e-02  2.58562006e-02 -5.86280152e-02\\n  1.86783802e-34 -2.57613063e-02 -1.05761262e-02 -1.49240410e-02\\n -4.53298853e-04 -2.51624547e-02 -1.69092566e-02  6.15713373e-03\\n -2.15551518e-02 -7.37830848e-02 -1.16919447e-02 -3.17851524e-03]\"\n        ],\n        \"semantic_type\": \"\",\n        \"description\": \"\"\n      }\n    }\n  ]\n}"
            }
          },
          "metadata": {},
          "execution_count": 28
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "import random\n",
        "\n",
        "import torch\n",
        "import numpy as np\n",
        "import pandas as pd\n",
        "\n",
        "device = \"cuda\" if torch.cuda.is_available() else \"cpu\"\n",
        "\n",
        "# Import texts and embedding df\n",
        "text_chunks_and_embedding_df = pd.read_csv(\"text_chunks_and_embeddings_df.csv\")\n",
        "\n",
        "# Convert embedding column back to np.array (it got converted to string when it got saved to CSV)\n",
        "text_chunks_and_embedding_df[\"embedding\"] = text_chunks_and_embedding_df[\"embedding\"].apply(lambda x: np.fromstring(x.strip(\"[]\"), sep=\" \"))\n",
        "\n",
        "# Convert texts and embedding df to list of dicts\n",
        "pages_and_chunks = text_chunks_and_embedding_df.to_dict(orient=\"records\")\n",
        "\n",
        "# Convert embeddings to torch tensor and send to device (note: NumPy arrays are float64, torch tensors are float32 by default)\n",
        "embeddings = torch.tensor(np.array(text_chunks_and_embedding_df[\"embedding\"].tolist()), dtype=torch.float32).to(device)\n",
        "embeddings.shape"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "3pULt1shkfnu",
        "outputId": "05a8fc19-07a0-4a62-b383-07ceecbdc445"
      },
      "execution_count": 29,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "torch.Size([35, 768])"
            ]
          },
          "metadata": {},
          "execution_count": 29
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "text_chunks_and_embedding_df.head()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 379
        },
        "id": "owxNhZMAkvKX",
        "outputId": "8f0c9b35-be47-4cc9-bd22-490d3e5e9671"
      },
      "execution_count": 30,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "   page_number                                     sentence_chunk  \\\n",
              "0          -41  REAL-TIME SEMANTIC SEGMENTATION ON FPGAS FOR A...   \n",
              "1          -40  challenges faced by, for example, the automoti...   \n",
              "2          -40  By applying aggressive ﬁlter-reduction and qua...   \n",
              "3          -40  With these steps, we obtain a good balance bet...   \n",
              "4          -40  In this way all inputs are smaller than one an...   \n",
              "\n",
              "   chunk_char_count  chunk_word_count  chunk_token_count  \\\n",
              "0              2364               329             591.00   \n",
              "1              1816               256             454.00   \n",
              "2              1633               234             408.25   \n",
              "3              1188               190             297.00   \n",
              "4               707               129             176.75   \n",
              "\n",
              "                                           embedding  \n",
              "0  [-0.0199189372, 0.0227944255, 0.00312705827, 0...  \n",
              "1  [-0.00901952107, 0.0335726589, -0.0203959588, ...  \n",
              "2  [-0.0413998924, 0.0438373275, -0.0120260157, 0...  \n",
              "3  [-0.00995968096, 0.0490480065, 0.036491558, 0....  \n",
              "4  [-0.0682296976, -0.0308206026, 0.0160923079, 0...  "
            ],
            "text/html": [
              "\n",
              "  <div id=\"df-e45acb49-3922-41ab-8f6b-4ba2e80959b6\" class=\"colab-df-container\">\n",
              "    <div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>page_number</th>\n",
              "      <th>sentence_chunk</th>\n",
              "      <th>chunk_char_count</th>\n",
              "      <th>chunk_word_count</th>\n",
              "      <th>chunk_token_count</th>\n",
              "      <th>embedding</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>-41</td>\n",
              "      <td>REAL-TIME SEMANTIC SEGMENTATION ON FPGAS FOR A...</td>\n",
              "      <td>2364</td>\n",
              "      <td>329</td>\n",
              "      <td>591.00</td>\n",
              "      <td>[-0.0199189372, 0.0227944255, 0.00312705827, 0...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>-40</td>\n",
              "      <td>challenges faced by, for example, the automoti...</td>\n",
              "      <td>1816</td>\n",
              "      <td>256</td>\n",
              "      <td>454.00</td>\n",
              "      <td>[-0.00901952107, 0.0335726589, -0.0203959588, ...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>-40</td>\n",
              "      <td>By applying aggressive ﬁlter-reduction and qua...</td>\n",
              "      <td>1633</td>\n",
              "      <td>234</td>\n",
              "      <td>408.25</td>\n",
              "      <td>[-0.0413998924, 0.0438373275, -0.0120260157, 0...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>-40</td>\n",
              "      <td>With these steps, we obtain a good balance bet...</td>\n",
              "      <td>1188</td>\n",
              "      <td>190</td>\n",
              "      <td>297.00</td>\n",
              "      <td>[-0.00995968096, 0.0490480065, 0.036491558, 0....</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>-40</td>\n",
              "      <td>In this way all inputs are smaller than one an...</td>\n",
              "      <td>707</td>\n",
              "      <td>129</td>\n",
              "      <td>176.75</td>\n",
              "      <td>[-0.0682296976, -0.0308206026, 0.0160923079, 0...</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>\n",
              "    <div class=\"colab-df-buttons\">\n",
              "\n",
              "  <div class=\"colab-df-container\">\n",
              "    <button class=\"colab-df-convert\" onclick=\"convertToInteractive('df-e45acb49-3922-41ab-8f6b-4ba2e80959b6')\"\n",
              "            title=\"Convert this dataframe to an interactive table.\"\n",
              "            style=\"display:none;\">\n",
              "\n",
              "  <svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\" viewBox=\"0 -960 960 960\">\n",
              "    <path d=\"M120-120v-720h720v720H120Zm60-500h600v-160H180v160Zm220 220h160v-160H400v160Zm0 220h160v-160H400v160ZM180-400h160v-160H180v160Zm440 0h160v-160H620v160ZM180-180h160v-160H180v160Zm440 0h160v-160H620v160Z\"/>\n",
              "  </svg>\n",
              "    </button>\n",
              "\n",
              "  <style>\n",
              "    .colab-df-container {\n",
              "      display:flex;\n",
              "      gap: 12px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert {\n",
              "      background-color: #E8F0FE;\n",
              "      border: none;\n",
              "      border-radius: 50%;\n",
              "      cursor: pointer;\n",
              "      display: none;\n",
              "      fill: #1967D2;\n",
              "      height: 32px;\n",
              "      padding: 0 0 0 0;\n",
              "      width: 32px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert:hover {\n",
              "      background-color: #E2EBFA;\n",
              "      box-shadow: 0px 1px 2px rgba(60, 64, 67, 0.3), 0px 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
              "      fill: #174EA6;\n",
              "    }\n",
              "\n",
              "    .colab-df-buttons div {\n",
              "      margin-bottom: 4px;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert {\n",
              "      background-color: #3B4455;\n",
              "      fill: #D2E3FC;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert:hover {\n",
              "      background-color: #434B5C;\n",
              "      box-shadow: 0px 1px 3px 1px rgba(0, 0, 0, 0.15);\n",
              "      filter: drop-shadow(0px 1px 2px rgba(0, 0, 0, 0.3));\n",
              "      fill: #FFFFFF;\n",
              "    }\n",
              "  </style>\n",
              "\n",
              "    <script>\n",
              "      const buttonEl =\n",
              "        document.querySelector('#df-e45acb49-3922-41ab-8f6b-4ba2e80959b6 button.colab-df-convert');\n",
              "      buttonEl.style.display =\n",
              "        google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
              "\n",
              "      async function convertToInteractive(key) {\n",
              "        const element = document.querySelector('#df-e45acb49-3922-41ab-8f6b-4ba2e80959b6');\n",
              "        const dataTable =\n",
              "          await google.colab.kernel.invokeFunction('convertToInteractive',\n",
              "                                                    [key], {});\n",
              "        if (!dataTable) return;\n",
              "\n",
              "        const docLinkHtml = 'Like what you see? Visit the ' +\n",
              "          '<a target=\"_blank\" href=https://colab.research.google.com/notebooks/data_table.ipynb>data table notebook</a>'\n",
              "          + ' to learn more about interactive tables.';\n",
              "        element.innerHTML = '';\n",
              "        dataTable['output_type'] = 'display_data';\n",
              "        await google.colab.output.renderOutput(dataTable, element);\n",
              "        const docLink = document.createElement('div');\n",
              "        docLink.innerHTML = docLinkHtml;\n",
              "        element.appendChild(docLink);\n",
              "      }\n",
              "    </script>\n",
              "  </div>\n",
              "\n",
              "\n",
              "<div id=\"df-edd5fa5e-0e6d-4942-9479-1e49b80763df\">\n",
              "  <button class=\"colab-df-quickchart\" onclick=\"quickchart('df-edd5fa5e-0e6d-4942-9479-1e49b80763df')\"\n",
              "            title=\"Suggest charts\"\n",
              "            style=\"display:none;\">\n",
              "\n",
              "<svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\"viewBox=\"0 0 24 24\"\n",
              "     width=\"24px\">\n",
              "    <g>\n",
              "        <path d=\"M19 3H5c-1.1 0-2 .9-2 2v14c0 1.1.9 2 2 2h14c1.1 0 2-.9 2-2V5c0-1.1-.9-2-2-2zM9 17H7v-7h2v7zm4 0h-2V7h2v10zm4 0h-2v-4h2v4z\"/>\n",
              "    </g>\n",
              "</svg>\n",
              "  </button>\n",
              "\n",
              "<style>\n",
              "  .colab-df-quickchart {\n",
              "      --bg-color: #E8F0FE;\n",
              "      --fill-color: #1967D2;\n",
              "      --hover-bg-color: #E2EBFA;\n",
              "      --hover-fill-color: #174EA6;\n",
              "      --disabled-fill-color: #AAA;\n",
              "      --disabled-bg-color: #DDD;\n",
              "  }\n",
              "\n",
              "  [theme=dark] .colab-df-quickchart {\n",
              "      --bg-color: #3B4455;\n",
              "      --fill-color: #D2E3FC;\n",
              "      --hover-bg-color: #434B5C;\n",
              "      --hover-fill-color: #FFFFFF;\n",
              "      --disabled-bg-color: #3B4455;\n",
              "      --disabled-fill-color: #666;\n",
              "  }\n",
              "\n",
              "  .colab-df-quickchart {\n",
              "    background-color: var(--bg-color);\n",
              "    border: none;\n",
              "    border-radius: 50%;\n",
              "    cursor: pointer;\n",
              "    display: none;\n",
              "    fill: var(--fill-color);\n",
              "    height: 32px;\n",
              "    padding: 0;\n",
              "    width: 32px;\n",
              "  }\n",
              "\n",
              "  .colab-df-quickchart:hover {\n",
              "    background-color: var(--hover-bg-color);\n",
              "    box-shadow: 0 1px 2px rgba(60, 64, 67, 0.3), 0 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
              "    fill: var(--button-hover-fill-color);\n",
              "  }\n",
              "\n",
              "  .colab-df-quickchart-complete:disabled,\n",
              "  .colab-df-quickchart-complete:disabled:hover {\n",
              "    background-color: var(--disabled-bg-color);\n",
              "    fill: var(--disabled-fill-color);\n",
              "    box-shadow: none;\n",
              "  }\n",
              "\n",
              "  .colab-df-spinner {\n",
              "    border: 2px solid var(--fill-color);\n",
              "    border-color: transparent;\n",
              "    border-bottom-color: var(--fill-color);\n",
              "    animation:\n",
              "      spin 1s steps(1) infinite;\n",
              "  }\n",
              "\n",
              "  @keyframes spin {\n",
              "    0% {\n",
              "      border-color: transparent;\n",
              "      border-bottom-color: var(--fill-color);\n",
              "      border-left-color: var(--fill-color);\n",
              "    }\n",
              "    20% {\n",
              "      border-color: transparent;\n",
              "      border-left-color: var(--fill-color);\n",
              "      border-top-color: var(--fill-color);\n",
              "    }\n",
              "    30% {\n",
              "      border-color: transparent;\n",
              "      border-left-color: var(--fill-color);\n",
              "      border-top-color: var(--fill-color);\n",
              "      border-right-color: var(--fill-color);\n",
              "    }\n",
              "    40% {\n",
              "      border-color: transparent;\n",
              "      border-right-color: var(--fill-color);\n",
              "      border-top-color: var(--fill-color);\n",
              "    }\n",
              "    60% {\n",
              "      border-color: transparent;\n",
              "      border-right-color: var(--fill-color);\n",
              "    }\n",
              "    80% {\n",
              "      border-color: transparent;\n",
              "      border-right-color: var(--fill-color);\n",
              "      border-bottom-color: var(--fill-color);\n",
              "    }\n",
              "    90% {\n",
              "      border-color: transparent;\n",
              "      border-bottom-color: var(--fill-color);\n",
              "    }\n",
              "  }\n",
              "</style>\n",
              "\n",
              "  <script>\n",
              "    async function quickchart(key) {\n",
              "      const quickchartButtonEl =\n",
              "        document.querySelector('#' + key + ' button');\n",
              "      quickchartButtonEl.disabled = true;  // To prevent multiple clicks.\n",
              "      quickchartButtonEl.classList.add('colab-df-spinner');\n",
              "      try {\n",
              "        const charts = await google.colab.kernel.invokeFunction(\n",
              "            'suggestCharts', [key], {});\n",
              "      } catch (error) {\n",
              "        console.error('Error during call to suggestCharts:', error);\n",
              "      }\n",
              "      quickchartButtonEl.classList.remove('colab-df-spinner');\n",
              "      quickchartButtonEl.classList.add('colab-df-quickchart-complete');\n",
              "    }\n",
              "    (() => {\n",
              "      let quickchartButtonEl =\n",
              "        document.querySelector('#df-edd5fa5e-0e6d-4942-9479-1e49b80763df button');\n",
              "      quickchartButtonEl.style.display =\n",
              "        google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
              "    })();\n",
              "  </script>\n",
              "</div>\n",
              "\n",
              "    </div>\n",
              "  </div>\n"
            ],
            "application/vnd.google.colaboratory.intrinsic+json": {
              "type": "dataframe",
              "variable_name": "text_chunks_and_embedding_df",
              "summary": "{\n  \"name\": \"text_chunks_and_embedding_df\",\n  \"rows\": 35,\n  \"fields\": [\n    {\n      \"column\": \"page_number\",\n      \"properties\": {\n        \"dtype\": \"number\",\n        \"std\": 3,\n        \"min\": -41,\n        \"max\": -31,\n        \"num_unique_values\": 11,\n        \"samples\": [\n          -36,\n          -41,\n          -32\n        ],\n        \"semantic_type\": \"\",\n        \"description\": \"\"\n      }\n    },\n    {\n      \"column\": \"sentence_chunk\",\n      \"properties\": {\n        \"dtype\": \"string\",\n        \"num_unique_values\": 35,\n        \"samples\": [\n          \"M. P., M. R., S. S., and V. L. are supported by the European Research Council (ERC) under the European Union\\u2019s Horizon 2020 research and innovation program (Grant Agreement No.772369). M. P. and N. G. are supported by the European Research Council (ERC) under the European Union\\u2019s Horizon 2020 research and innovation program (Grant Agreement No.966696).10\",\n          \"In our case, we use AUTOQKERAS [12]. In AUTOQKERAS, a hyperparameter search over individual layer quantization conditions and \\ufb01lter counts is performed. Since the model contains skip connections, the scan over number of \\ufb01lters needs to be handled with care. In particular, we use the block features of AUTOQKERAS to ensure that the \\ufb01lter count matches throughout a bottleneck, so that the tensor addition of the skip connection will have valid dimensions. The search for best hyperparameters, including the choice of indivdual quantizers for kernels and activations, is carried out using a Bayesian strategy where the balance between accuracy and resource usage is controlled by targeting a metric derived from them both[12]. In our search we permit e.g. a 4% decrease in accuracy if the resource usage also is halved at the same time.6\",\n          \"Including these, the average time per image increases to 720 ms.6 Conclusions In this paper, we demonstrate that we can perform semantic segmentation on a single FPGA on a Zynq MPSoC device using a compressed version of ENet. The network is compressed using automatic heterogeneous quantization at training time and a \\ufb01lter ablation procedure, and is then evaluated on the Cityscapes dataset. Inference is executed on hardware with a latency of 4.9 ms per image, utilizing 18% of the DSPs, 30% of the LUTs, 16% of the FFs and 25 % of the BRAMs. Processing the images in batches of ten results in a latency of 30 ms per batch, which is signi\\ufb01cantly faster than ten times the single-image batch inference latency. This is relevant when batching over images captured from different cameras simultaneously. By introducing an improved implementation of convolutional layers in hls4ml, we signi\\ufb01cantly reduce resource consumption, allowing for a fully-on-chip deployment of larger convolutional neural networks. This avoids latency overhead caused by data transfers between off-chip memory and FPGA processing elements, or between multiple devices. Also taking into account the favorable power-ef\\ufb01ciency of FPGAs, we conclude that FPGAs offer highly interesting, low-power alternatives to GPUs for on-vehicle deep learning inference and other computer vision tasks requiring low-power and low-latency. Acknowledgment We acknowledge the Fast Machine Learning collective as an open community of multi-domain experts and collaborators.\"\n        ],\n        \"semantic_type\": \"\",\n        \"description\": \"\"\n      }\n    },\n    {\n      \"column\": \"chunk_char_count\",\n      \"properties\": {\n        \"dtype\": \"number\",\n        \"std\": 590,\n        \"min\": 66,\n        \"max\": 2364,\n        \"num_unique_values\": 34,\n        \"samples\": [\n          1152,\n          992,\n          517\n        ],\n        \"semantic_type\": \"\",\n        \"description\": \"\"\n      }\n    },\n    {\n      \"column\": \"chunk_word_count\",\n      \"properties\": {\n        \"dtype\": \"number\",\n        \"std\": 99,\n        \"min\": 10,\n        \"max\": 425,\n        \"num_unique_values\": 35,\n        \"samples\": [\n          56,\n          134,\n          229\n        ],\n        \"semantic_type\": \"\",\n        \"description\": \"\"\n      }\n    },\n    {\n      \"column\": \"chunk_token_count\",\n      \"properties\": {\n        \"dtype\": \"number\",\n        \"std\": 147.61663313512898,\n        \"min\": 16.5,\n        \"max\": 591.0,\n        \"num_unique_values\": 34,\n        \"samples\": [\n          288.0,\n          248.0,\n          129.25\n        ],\n        \"semantic_type\": \"\",\n        \"description\": \"\"\n      }\n    },\n    {\n      \"column\": \"embedding\",\n      \"properties\": {\n        \"dtype\": \"object\",\n        \"semantic_type\": \"\",\n        \"description\": \"\"\n      }\n    }\n  ]\n}"
            }
          },
          "metadata": {},
          "execution_count": 30
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "from sentence_transformers import util, SentenceTransformer\n",
        "\n",
        "embedding_model = SentenceTransformer(model_name_or_path=\"all-mpnet-base-v2\",\n",
        "                                      device=device) # choose the device to load the model to"
      ],
      "metadata": {
        "id": "4-zzrTtLk5k0"
      },
      "execution_count": 31,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "# Searching"
      ],
      "metadata": {
        "id": "F2Dwt9Gwlbzi"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "# 1. Define the query\n",
        "# Note: This could be anything. But since we're working with a nutrition textbook, we'll stick with nutrition-based queries.\n",
        "query = \"fully-on-chip deployment with a latency\"\n",
        "print(f\"Query: {query}\")\n",
        "\n",
        "# 2. Embed the query to the same numerical space as the text examples\n",
        "# Note: It's important to embed your query with the same model you embedded your examples with.\n",
        "query_embedding = embedding_model.encode(query, convert_to_tensor=True)\n",
        "\n",
        "# 3. Get similarity scores with the dot product (we'll time this for fun)\n",
        "from time import perf_counter as timer\n",
        "\n",
        "start_time = timer()\n",
        "dot_scores = util.dot_score(a=query_embedding, b=embeddings)[0]\n",
        "end_time = timer()\n",
        "\n",
        "print(f\"Time take to get scores on {len(embeddings)} embeddings: {end_time-start_time:.5f} seconds.\")\n",
        "\n",
        "# 4. Get the top-k results (we'll keep this to 5)\n",
        "top_results_dot_product = torch.topk(dot_scores, k=5)\n",
        "top_results_dot_product"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "YxP8y2_IldgP",
        "outputId": "fc13c431-666a-4a95-eca8-a7ad326fa1ee"
      },
      "execution_count": 32,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Query: fully-on-chip deployment with a latency\n",
            "Time take to get scores on 35 embeddings: 0.00296 seconds.\n"
          ]
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "torch.return_types.topk(\n",
              "values=tensor([0.5494, 0.4831, 0.4765, 0.3932, 0.3429], device='cuda:0'),\n",
              "indices=tensor([23, 19, 18, 21, 20], device='cuda:0'))"
            ]
          },
          "metadata": {},
          "execution_count": 32
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "larger_embeddings = torch.randn(100*embeddings.shape[0], 768).to(device)\n",
        "print(f\"Embeddings shape: {larger_embeddings.shape}\")\n",
        "\n",
        "# Perform dot product across 168,000 embeddings\n",
        "start_time = timer()\n",
        "dot_scores = util.dot_score(a=query_embedding, b=larger_embeddings)[0]\n",
        "end_time = timer()\n",
        "\n",
        "print(f\"Time take to get scores on {len(larger_embeddings)} embeddings: {end_time-start_time:.5f} seconds.\")"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "ku_RVpsmmLaM",
        "outputId": "b04eddbd-37ba-47c0-dad6-a3a5d88fb52e"
      },
      "execution_count": 33,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Embeddings shape: torch.Size([3500, 768])\n",
            "Time take to get scores on 3500 embeddings: 0.02678 seconds.\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# Define helper function to print wrapped text\n",
        "import textwrap\n",
        "\n",
        "def print_wrapped(text, wrap_length=80):\n",
        "    wrapped_text = textwrap.fill(text, wrap_length)\n",
        "    print(wrapped_text)"
      ],
      "metadata": {
        "id": "jWgQ1rq-mSx8"
      },
      "execution_count": 34,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "print(f\"Query: '{query}'\\n\")\n",
        "print(\"Results:\")\n",
        "# Loop through zipped together scores and indicies from torch.topk\n",
        "for score, idx in zip(top_results_dot_product[0], top_results_dot_product[1]):\n",
        "    print(f\"Score: {score:.4f}\")\n",
        "    # Print relevant sentence chunk (since the scores are in descending order, the most relevant chunk will be first)\n",
        "    print(\"Text:\")\n",
        "    print_wrapped(pages_and_chunks[idx][\"sentence_chunk\"])\n",
        "    # Print the page number too so we can reference the textbook further (and check the results)\n",
        "    print(f\"Page number: {pages_and_chunks[idx]['page_number']}\")\n",
        "    print(\"\\n\")"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "k24YK7ZOmXJ1",
        "outputId": "388cdc2f-3afd-46db-e4ca-dbdf2871c6e2"
      },
      "execution_count": 35,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Query: 'fully-on-chip deployment with a latency'\n",
            "\n",
            "Results:\n",
            "Score: 0.5494\n",
            "Text:\n",
            "In order to achieve the lowest possible latency, we implement a fully on-chip\n",
            "design with high layer parallelism. We optimize for latency, rather than frame\n",
            "rate, such that in a real-life application the vehicle response time could be\n",
            "minimized. Keeping up with the camera frame rate is a minimal requirement, but a\n",
            "latency lower than the frame interval can be utilized. In our approach, each\n",
            "layer is implemented as a separate module and data is streamed through the\n",
            "architecture layer by layer. Dedicated per-layer buffers ensure that just enough\n",
            "data is buffered in order to feed the next layer. This is highly efﬁcient, but\n",
            "limits the number of layers that can be implemented on the FPGA. Consequently,\n",
            "in order to ﬁt onto the FPGA in question, our model is smaller and achieves a\n",
            "lower mIoU. Ref. [23] does not quote a latency, but a frame rate. A best-case\n",
            "latency is then computed as the inverse of this frame rate, which corresponds to\n",
            "30.38 ms. However, this does not include any overhead latency like data\n",
            "transfer, pre- and post-processing.\n",
            "Page number: -33\n",
            "\n",
            "\n",
            "Score: 0.4831\n",
            "Text:\n",
            "5.3 Results The hardware we target is a Zynq UltraScale+ MPSoC device\n",
            "(xczu9eg-2ffvb1156) on a ZCU102 development kit, which targets automotive\n",
            "applications. After reducing the FPGA resource consumption through the methods\n",
            "described above, the highest accuracy models highlighted in Table 3 are\n",
            "synthesized. These are the homogeneously quantized Enet8Q8 and Enet8Q4 models,\n",
            "as well as the heterogeneously quantized EnetHQ model. To ﬁnd the lowest latency\n",
            "implementation, we run several attempts varying the reuse factor (RF) and the\n",
            "clock period. The RF indicates how many times a multiplier can be reused (zero\n",
            "for a fully parallel implementation). Lower RF leads to lower latency, but\n",
            "higher resource usage. We targeted reuse factors of 10, 20, 50, 100, and clock\n",
            "periods of 5, 7, 10 ns. For each model, we then chose the conﬁguration yielding\n",
            "the lowest latency. For Enet8Q8, this is a target clock period of 7 ns and\n",
            "RF=10. For Enet8Q4 and EnetHQ we use a clock period of 7 ns and RF = 6.\n",
            "Page number: -34\n",
            "\n",
            "\n",
            "Score: 0.4765\n",
            "Text:\n",
            "By using the clock-cycle accurate RTL simulation of the scheduled design, we can\n",
            "monitor the actual occupancy of each FIFO in the model when running the\n",
            "simulation over example images. This enables us to extract and set the correct\n",
            "size of the FIFOs, reducing memory usage compared to the baseline. By applying\n",
            "this procedure, we observe a memory efﬁciency P l Ol P l Fl = 19.5%, where the\n",
            "index l runs across the layers, Ol is the observed occupancy for the l-th layer,\n",
            "and Fl is the corresponding FIFO depth. The corresponding mean occupancy is\n",
            "found to be P l Ol Fl = 4.7%. We then resize every FIFO to its observed maximum\n",
            "occupancy and rerun the C-Synthesis, thereby saving FPGA resources and allowing\n",
            "larger models to ﬁt on the FPGA. Table 5 shows the impact of such an\n",
            "optimization on the FPGA resources for one example model, Enet8Q8, demonstrating\n",
            "a signiﬁcant reduction of resources. Optimisation BRAM LUT FF DSP Latency No\n",
            "7,270 676,760 230,913 228 3.577 ms Yes 1,398 437,559 146,392 228 3.577 ms\n",
            "Improvement -81% -35% -37% 0% 0% Table 5: Effect of FIFO depth optimization on\n",
            "FPGA resource usage and model latency. The values in the table are taken from\n",
            "Vivado HLS estimates of resource usage. A comparison using physical resource\n",
            "usage is unfeasible since the model without optimization can not be synthesized.\n",
            "The estimates are made targeting an xczu9eg-2ffvb1156 MPSoC device on a ZCU102\n",
            "development kit.\n",
            "Page number: -34\n",
            "\n",
            "\n",
            "Score: 0.3932\n",
            "Text:\n",
            "Table 6: Accuracy, mIoU, latency and resource utilization for the EnetHQ,\n",
            "Enet8Q4 and Enet8Q8 models. The latency is quoted for a batch size b = 1 and b =\n",
            "10. Resources are expressed as a percentage of those available on the\n",
            "xczu9eg-2ffvb1156 MPSoC device on the ZCU102 development kit. The last row is a\n",
            "comparison to work presented in Ref. [23]. Model Acc.mIoU Latency [ms] BRAM LUT\n",
            "FF DSP b=1 b=10 EnetHQ 81.1% 36.8 % 4.9 30.6 224.5 (25%) 76,718 (30%) 87,059\n",
            "(16%) 450 (18%) Enet8Q4 77.6% 33.9 % 4.8 30.2 342.0 (37%) 166,741 (61%) 90,536\n",
            "(16%) 0 Enet8Q8 77.1% 33.4 % 4.8 30.0 508.5 (56%) 126,458 (46%) 134,385 (25%)\n",
            "1,502 (60%) ENet [23] - 63.1% 30.38 (720)a - 257 62,599 192,212 689 a The former\n",
            "is without considering data transfer, pre- and post-processing. The number in\n",
            "parenthesis includes these additional overheads, averaged over 58 images, and is\n",
            "more comparable to the numbers we present. We measured the time taken by the\n",
            "accelerator to produce a prediction on batches of images, with batch sizes of b\n",
            "= 1 and b = 10. The same predictions have been executed 105 times, and the time\n",
            "average is taken as the latency.\n",
            "Page number: -33\n",
            "\n",
            "\n",
            "Score: 0.3429\n",
            "Text:\n",
            "Inference performance of this model was measured on the ZCU102 target device.\n",
            "The ﬁnal latency and resource utilization report is shown in Table 5.3.8\n",
            "Page number: -34\n",
            "\n",
            "\n"
          ]
        }
      ]
    }
  ]
}