.ipynb_checkpoints/random-sample-checkpoint.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 35,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.datasets import fetch_openml\n",
    "import matplotlib.pyplot as plt\n",
    "import itertools\n",
    "import numpy as np\n",
    "import pandas as pd\n",
    "from keras.utils.np_utils import to_categorical\n",
    "import tensorflow as tf\n",
    "from tensorflow import keras\n",
    "from tensorflow.keras import layers\n",
    "from tensorflow.keras.models import Sequential, load_model\n",
    "from keras.datasets import mnist\n",
    "import itertools\n",
    "plt.style.use('dark_background')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "metadata": {},
   "outputs": [],
   "source": [
    "(train_images, train_labels), (test_images, test_labels) = mnist.load_data()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(28, 28)\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "<matplotlib.image.AxesImage at 0x1553f83a0>"
      ]
     },
     "execution_count": 37,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "image_A = train_images[0]\n",
    "print(image_A.shape)\n",
    "plt.imshow(image_A, cmap = plt.cm.binary)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 38,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.image.AxesImage at 0x154e5a910>"
      ]
     },
     "execution_count": 38,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "image_B = train_images[1]\n",
    "plt.imshow(image_B, cmap = plt.cm.binary)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 39,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.image.AxesImage at 0x154eba0a0>"
      ]
     },
     "execution_count": 39,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "images_A_and_B = np.hstack((image_A, image_B))\n",
    "plt.imshow(images_A_and_B, cmap = plt.cm.binary)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 40,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Create mini training set so that it's not too computationally intensive\n",
    "mini_size = 500\n",
    "mini_X_train, mini_X_test, mini_y_train, mini_y_test = train_images[:mini_size], train_images[mini_size:mini_size*2],  train_labels[:mini_size], train_labels[mini_size:mini_size*2]\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 41,
   "metadata": {},
   "outputs": [],
   "source": [
    "num_images = mini_X_train.shape[0]\n",
    "index_list = np.linspace(0,num_images-1,num_images).astype(int)\n",
    "image_combinations = list(itertools.combinations(index_list, 2))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 42,
   "metadata": {},
   "outputs": [],
   "source": [
    "# First just try putting them one on top of the other\n",
    "concatenated_dataset_X = []\n",
    "concatenated_dataset_y = []\n",
    "\n",
    "for combination in image_combinations:\n",
    "#     print(combination)\n",
    "    \n",
    "    index_A = combination[0]\n",
    "    index_B = combination[1]\n",
    "    \n",
    "    image_A = mini_X_train[index_A]\n",
    "    image_B = mini_X_train[index_B]\n",
    "    \n",
    "    intermediate_label_A = mini_y_train[index_A]\n",
    "    intermediate_label_B = mini_y_train[index_B]\n",
    "    \n",
    "    concatenated_images = np.hstack((image_A, image_B))\n",
    "    \n",
    "    final_label = min(intermediate_label_A, intermediate_label_B)\n",
    "    \n",
    "    concatenated_dataset_X.append(concatenated_images)\n",
    "    concatenated_dataset_y.append(final_label)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 43,
   "metadata": {},
   "outputs": [],
   "source": [
    "num_images = mini_X_test.shape[0]\n",
    "index_list = np.linspace(0,num_images-1,num_images).astype(int)\n",
    "image_combinations = list(itertools.combinations(index_list, 2))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 44,
   "metadata": {},
   "outputs": [],
   "source": [
    "# As above for test\n",
    "concatenated_dataset_X_test = []\n",
    "concatenated_dataset_y_test = []\n",
    "\n",
    "for combination in image_combinations:\n",
    "#     print(combination)\n",
    "    \n",
    "    index_A = combination[0]\n",
    "    index_B = combination[1]\n",
    "    \n",
    "    image_A = mini_X_test[index_A]\n",
    "    image_B = mini_X_test[index_B]\n",
    "    \n",
    "    intermediate_label_A = mini_y_test[index_A]\n",
    "    intermediate_label_B = mini_y_test[index_B]\n",
    "    \n",
    "    concatenated_images = np.hstack((image_A, image_B))\n",
    "    \n",
    "    final_label = min(intermediate_label_A, intermediate_label_B)\n",
    "    \n",
    "    concatenated_dataset_X_test.append(concatenated_images)\n",
    "    concatenated_dataset_y_test.append(final_label)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 45,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(124750, 28, 56)\n",
      "(124750, 28, 56)\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "124750"
      ]
     },
     "execution_count": 45,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# To array\n",
    "concatenated_dataset_X = np.array(concatenated_dataset_X)\n",
    "concatenated_dataset_y = np.array(concatenated_dataset_y)\n",
    "\n",
    "concatenated_dataset_X_test = np.array(concatenated_dataset_X_test)\n",
    "concatenated_dataset_y_test = np.array(concatenated_dataset_y_test)\n",
    "\n",
    "print(concatenated_dataset_X.shape)\n",
    "print(concatenated_dataset_X_test.shape)\n",
    "\n",
    "input_size = concatenated_dataset_X.shape[0]\n",
    "input_size"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 59,
   "metadata": {},
   "outputs": [],
   "source": [
    "train_images = concatenated_dataset_X.reshape((input_size, 28, 56, 1))\n",
    "train_images = train_images.astype('float32')/ 255\n",
    "\n",
    "test_images = concatenated_dataset_X_test.reshape((input_size, 28, 56, 1))\n",
    "test_images = test_images.astype('float32')/255\n",
    "\n",
    "train_labels = to_categorical(concatenated_dataset_y)\n",
    "test_labels = to_categorical(concatenated_dataset_y_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 60,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(124750, 28, 56, 1)\n",
      "(124750, 10)\n"
     ]
    }
   ],
   "source": [
    "# Check shapes\n",
    "print(train_images.shape)\n",
    "print(train_labels.shape)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 48,
   "metadata": {},
   "outputs": [],
   "source": [
    "model = tf.keras.Sequential()\n",
    "model.add(layers.Conv2D(32, (3,3), activation = 'relu', input_shape = (28,56,1)))\n",
    "model.add(layers.MaxPooling2D(2,2))\n",
    "model.add(layers.Conv2D(64, (3,3), activation = 'relu'))\n",
    "model.add(layers.MaxPooling2D(2,2))\n",
    "model.add(layers.Conv2D(64, (3,3), activation = 'relu'))\n",
    "# Adding a classifier\n",
    "model.add(layers.Flatten())\n",
    "model.add(layers.Dense(64, activation = 'relu'))\n",
    "model.add(layers.Dense(10, activation = 'softmax'))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 49,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Model: \"sequential_1\"\n",
      "_________________________________________________________________\n",
      "Layer (type)                 Output Shape              Param #   \n",
      "=================================================================\n",
      "conv2d_3 (Conv2D)            (None, 26, 54, 32)        320       \n",
      "_________________________________________________________________\n",
      "max_pooling2d_2 (MaxPooling2 (None, 13, 27, 32)        0         \n",
      "_________________________________________________________________\n",
      "conv2d_4 (Conv2D)            (None, 11, 25, 64)        18496     \n",
      "_________________________________________________________________\n",
      "max_pooling2d_3 (MaxPooling2 (None, 5, 12, 64)         0         \n",
      "_________________________________________________________________\n",
      "conv2d_5 (Conv2D)            (None, 3, 10, 64)         36928     \n",
      "_________________________________________________________________\n",
      "flatten_1 (Flatten)          (None, 1920)              0         \n",
      "_________________________________________________________________\n",
      "dense_2 (Dense)              (None, 64)                122944    \n",
      "_________________________________________________________________\n",
      "dense_3 (Dense)              (None, 10)                650       \n",
      "=================================================================\n",
      "Total params: 179,338\n",
      "Trainable params: 179,338\n",
      "Non-trainable params: 0\n",
      "_________________________________________________________________\n"
     ]
    }
   ],
   "source": [
    "model.summary()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 50,
   "metadata": {},
   "outputs": [],
   "source": [
    "model.compile(optimizer = 'rmsprop',\n",
    "             loss = 'categorical_crossentropy',\n",
    "             metrics = ['accuracy'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 51,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/5\n",
      "1950/1950 [==============================] - 90s 46ms/step - loss: 0.0842 - accuracy: 0.9729\n",
      "Epoch 2/5\n",
      "1950/1950 [==============================] - 93s 48ms/step - loss: 0.0021 - accuracy: 0.9994\n",
      "Epoch 3/5\n",
      "1950/1950 [==============================] - 94s 48ms/step - loss: 0.0017 - accuracy: 0.9996\n",
      "Epoch 4/5\n",
      "1950/1950 [==============================] - 96s 49ms/step - loss: 1.3761e-04 - accuracy: 1.0000\n",
      "Epoch 5/5\n",
      "1950/1950 [==============================] - 92s 47ms/step - loss: 4.2182e-04 - accuracy: 0.9999\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "<tensorflow.python.keras.callbacks.History at 0x1578b8a00>"
      ]
     },
     "execution_count": 51,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Run model in one go\n",
    "model.fit(train_images, train_labels, epochs = 5, batch_size = 64)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Noting that the accuracy is not changing much here"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 62,
   "metadata": {},
   "outputs": [],
   "source": [
    "# split test_images into a val set as vell\n",
    "val_images = test_images[0:200]\n",
    "val_labels = test_labels[0:200]\n",
    "\n",
    "test_images = test_images[200:0]\n",
    "test_labels = test_labels[200:0]\n",
    "\n",
    "# test_loss, test_acc = model.evaluate(test_images, test_labels)\n",
    "# test_acc"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 74,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[0. 0. 0. 1. 0. 0. 0. 0. 0. 0.]\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "img_num = 100\n",
    "plt.imshow(val_images[img_num])\n",
    "print(val_labels[img_num])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "It seems like I am going to need a random set of combinations"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 64,
   "metadata": {},
   "outputs": [],
   "source": [
    "# i = 0\n",
    "# val_data = train_images[i * num_val_samples: (i + 1) * num_val_samples]\n",
    "# val_data.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Set up parameters for cross-validation\n",
    "k = 3\n",
    "num_val_samples = len(train_images)//k"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "processing fold # 0\n",
      "processing fold # 1\n",
      "processing fold # 2\n",
      "processing fold # 3\n"
     ]
    }
   ],
   "source": [
    "# Try with more epochs\n",
    "\n",
    "num_epochs = 20\n",
    "all_mae_histories = []\n",
    "\n",
    "for i in range(k):\n",
    "    \n",
    "    print('processing fold #', i)\n",
    "    val_data = train_images[i * num_val_samples: (i + 1) * num_val_samples]\n",
    "    val_targets = train_labels[i * num_val_samples: (i + 1) * num_val_samples]\n",
    "    \n",
    "    partial_train_data = np.concatenate(\n",
    "    [train_images[:i * num_val_samples],\n",
    "    train_images[(i+1) * num_val_samples:]],\n",
    "    axis = 0)\n",
    "    \n",
    "    partial_train_targets = np.concatenate(\n",
    "    [train_labels[:i*num_val_samples],\n",
    "    train_labels[(i+1)*num_val_samples:]],\n",
    "    axis = 0)\n",
    "    \n",
    "    model.compile(optimizer = 'rmsprop',\n",
    "             loss = 'categorical_crossentropy',\n",
    "             metrics = ['accuracy', 'mae'])\n",
    "    \n",
    "    history = model.fit(partial_train_data,\n",
    "              partial_train_targets,\n",
    "              validation_data = (val_data, val_targets),\n",
    "              epochs = num_epochs,\n",
    "              batch_size = 64,\n",
    "              verbose = 0)\n",
    "    \n",
    "    val_mae_history = history.history['val_mae']\n",
    "    #val_acc_history = history.history['accuracy']\n",
    "    #print(history.history.keys())\n",
    "    all_mae_histories.append(val_mae_history)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Compute the average of the per-epoch MAE scores for all folds\n",
    "# Building the history of successive mean K-fold validation scores\n",
    "\n",
    "average_mae_history = [\n",
    "    np.mean([x[i] for x in all_mae_histories]) for i in range(num_epochs)]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.plot(range(1, len(average_mae_history) + 1), average_mae_history)\n",
    "plt.xlabel('Epochs')\n",
    "plt.ylabel('val_mae')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "metadata": {},
   "outputs": [],
   "source": [
    "def smooth_curve(points, factor =0.9):\n",
    "    \n",
    "    smoothed_points = []\n",
    "    \n",
    "    for point in points:\n",
    "        if smoothed_points:\n",
    "            previous = smoothed_points[-1]\n",
    "            smoothed_points.append(previous * factor + point * (1 - factor))\n",
    "        else:\n",
    "            smoothed_points.append(point)\n",
    "        \n",
    "    return smoothed_points\n",
    "\n",
    "smooth_mae_history = smooth_curve(average_mae_history)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 53,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Difficult to understand plot due to scaling issues\n",
    "plt.plot(range(1, len(smooth_mae_history) + 1), smooth_mae_history)\n",
    "plt.xlabel('Epochs')\n",
    "plt.ylabel('val_mae')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "general_ML",
   "language": "python",
   "name": "general_ml"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.9.2"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}