fix py38 and support python3.11

[marigoold]

No description provided.

[github-actions]

View latest API docs preview at: https://oneflow-staging.oss-cn-beijing.aliyuncs.com/docs/Oneflow-Inc/oneflow/pr/10391/

[github-actions]

View latest API docs preview at: https://oneflow-staging.oss-cn-beijing.aliyuncs.com/docs/Oneflow-Inc/oneflow/pr/10391/

[github-actions]

Speed stats:

GPU Name: NVIDIA GeForce RTX 3080 Ti 

❌ OneFlow resnet50 time: 43.8ms (= 4375.9ms / 100, input_shape=[16, 3, 224, 224])
PyTorch resnet50 time: 58.1ms (= 5810.0ms / 100, input_shape=[16, 3, 224, 224])
✔️ Relative speed: 1.33 (= 58.1ms / 43.8ms)

OneFlow resnet50 time: 26.7ms (= 2669.5ms / 100, input_shape=[8, 3, 224, 224])
PyTorch resnet50 time: 38.3ms (= 3829.6ms / 100, input_shape=[8, 3, 224, 224])
✔️ Relative speed: 1.43 (= 38.3ms / 26.7ms)

OneFlow resnet50 time: 18.3ms (= 3669.3ms / 200, input_shape=[4, 3, 224, 224])
PyTorch resnet50 time: 36.8ms (= 7358.4ms / 200, input_shape=[4, 3, 224, 224])
✔️ Relative speed: 2.01 (= 36.8ms / 18.3ms)

OneFlow resnet50 time: 17.2ms (= 3450.0ms / 200, input_shape=[2, 3, 224, 224])
PyTorch resnet50 time: 39.2ms (= 7840.8ms / 200, input_shape=[2, 3, 224, 224])
✔️ Relative speed: 2.27 (= 39.2ms / 17.2ms)

OneFlow resnet50 time: 18.2ms (= 3641.1ms / 200, input_shape=[1, 3, 224, 224])
PyTorch resnet50 time: 28.2ms (= 5638.5ms / 200, input_shape=[1, 3, 224, 224])
✔️ Relative speed: 1.55 (= 28.2ms / 18.2ms)

OneFlow swin dataloader time: 0.201s (= 40.261s / 200, num_workers=1)
PyTorch swin dataloader time: 0.128s (= 25.535s / 200, num_workers=1)
Relative speed: 0.634 (= 0.128s / 0.201s)

OneFlow swin dataloader time: 0.055s (= 11.093s / 200, num_workers=4)
PyTorch swin dataloader time: 0.032s (= 6.459s / 200, num_workers=4)
Relative speed: 0.582 (= 0.032s / 0.055s)

OneFlow swin dataloader time: 0.031s (= 6.246s / 200, num_workers=8)
PyTorch swin dataloader time: 0.017s (= 3.412s / 200, num_workers=8)
Relative speed: 0.546 (= 0.017s / 0.031s)

❌ OneFlow resnet50 time: 52.3ms (= 5233.7ms / 100, input_shape=[16, 3, 224, 224], ddp, world size=2)
PyTorch resnet50 time: 68.1ms (= 6814.9ms / 100, input_shape=[16, 3, 224, 224], ddp, world size=2)
✔️ Relative speed: 1.30 (= 68.1ms / 52.3ms)

OneFlow resnet50 time: 36.6ms (= 3662.4ms / 100, input_shape=[8, 3, 224, 224], ddp, world size=2)
PyTorch resnet50 time: 46.7ms (= 4665.8ms / 100, input_shape=[8, 3, 224, 224], ddp, world size=2)
✔️ Relative speed: 1.27 (= 46.7ms / 36.6ms)

OneFlow resnet50 time: 29.3ms (= 5857.9ms / 200, input_shape=[4, 3, 224, 224], ddp, world size=2)
PyTorch resnet50 time: 43.0ms (= 8605.6ms / 200, input_shape=[4, 3, 224, 224], ddp, world size=2)
✔️ Relative speed: 1.47 (= 43.0ms / 29.3ms)

OneFlow resnet50 time: 26.6ms (= 5329.0ms / 200, input_shape=[2, 3, 224, 224], ddp, world size=2)
PyTorch resnet50 time: 38.4ms (= 7674.6ms / 200, input_shape=[2, 3, 224, 224], ddp, world size=2)
✔️ Relative speed: 1.44 (= 38.4ms / 26.6ms)

OneFlow resnet50 time: 24.0ms (= 4797.0ms / 200, input_shape=[1, 3, 224, 224], ddp, world size=2)
PyTorch resnet50 time: 37.3ms (= 7467.2ms / 200, input_shape=[1, 3, 224, 224], ddp, world size=2)
✔️ Relative speed: 1.56 (= 37.3ms / 24.0ms)