fix release cu118 pkg

[jackalcooper]

No description provided.

[github-actions]

View latest API docs preview at: https://staging.oneflow.info/docs/Oneflow-Inc/oneflow/pr/10325/

[github-actions]

Speed stats:

GPU Name: NVIDIA GeForce RTX 3080 Ti 

❌ OneFlow resnet50 time: 43.6ms (= 4364.6ms / 100, input_shape=[16, 3, 224, 224])
PyTorch resnet50 time: 57.4ms (= 5743.2ms / 100, input_shape=[16, 3, 224, 224])
✔️ Relative speed: 1.32 (= 57.4ms / 43.6ms)

OneFlow resnet50 time: 26.0ms (= 2603.0ms / 100, input_shape=[8, 3, 224, 224])
PyTorch resnet50 time: 37.7ms (= 3774.4ms / 100, input_shape=[8, 3, 224, 224])
✔️ Relative speed: 1.45 (= 37.7ms / 26.0ms)

OneFlow resnet50 time: 19.5ms (= 3892.9ms / 200, input_shape=[4, 3, 224, 224])
PyTorch resnet50 time: 36.0ms (= 7202.4ms / 200, input_shape=[4, 3, 224, 224])
✔️ Relative speed: 1.85 (= 36.0ms / 19.5ms)

OneFlow resnet50 time: 18.1ms (= 3629.4ms / 200, input_shape=[2, 3, 224, 224])
PyTorch resnet50 time: 31.1ms (= 6216.6ms / 200, input_shape=[2, 3, 224, 224])
✔️ Relative speed: 1.71 (= 31.1ms / 18.1ms)

OneFlow resnet50 time: 18.2ms (= 3648.3ms / 200, input_shape=[1, 3, 224, 224])
PyTorch resnet50 time: 30.0ms (= 6005.2ms / 200, input_shape=[1, 3, 224, 224])
✔️ Relative speed: 1.65 (= 30.0ms / 18.2ms)

OneFlow swin dataloader time: 0.200s (= 39.984s / 200, num_workers=1)
PyTorch swin dataloader time: 0.129s (= 25.863s / 200, num_workers=1)
Relative speed: 0.647 (= 0.129s / 0.200s)

OneFlow swin dataloader time: 0.053s (= 10.652s / 200, num_workers=4)
PyTorch swin dataloader time: 0.033s (= 6.646s / 200, num_workers=4)
Relative speed: 0.624 (= 0.033s / 0.053s)

OneFlow swin dataloader time: 0.030s (= 6.086s / 200, num_workers=8)
PyTorch swin dataloader time: 0.016s (= 3.261s / 200, num_workers=8)
Relative speed: 0.536 (= 0.016s / 0.030s)

❌ OneFlow resnet50 time: 47.7ms (= 4768.9ms / 100, input_shape=[16, 3, 224, 224], ddp, world size=2)
PyTorch resnet50 time: 63.8ms (= 6382.4ms / 100, input_shape=[16, 3, 224, 224], ddp, world size=2)
✔️ Relative speed: 1.34 (= 63.8ms / 47.7ms)

OneFlow resnet50 time: 31.0ms (= 3101.7ms / 100, input_shape=[8, 3, 224, 224], ddp, world size=2)
PyTorch resnet50 time: 43.8ms (= 4378.6ms / 100, input_shape=[8, 3, 224, 224], ddp, world size=2)
✔️ Relative speed: 1.41 (= 43.8ms / 31.0ms)

OneFlow resnet50 time: 24.0ms (= 4798.8ms / 200, input_shape=[4, 3, 224, 224], ddp, world size=2)
PyTorch resnet50 time: 41.5ms (= 8294.2ms / 200, input_shape=[4, 3, 224, 224], ddp, world size=2)
✔️ Relative speed: 1.73 (= 41.5ms / 24.0ms)

OneFlow resnet50 time: 21.5ms (= 4306.7ms / 200, input_shape=[2, 3, 224, 224], ddp, world size=2)
PyTorch resnet50 time: 38.4ms (= 7671.1ms / 200, input_shape=[2, 3, 224, 224], ddp, world size=2)
✔️ Relative speed: 1.78 (= 38.4ms / 21.5ms)

OneFlow resnet50 time: 20.9ms (= 4170.6ms / 200, input_shape=[1, 3, 224, 224], ddp, world size=2)
PyTorch resnet50 time: 34.1ms (= 6821.8ms / 200, input_shape=[1, 3, 224, 224], ddp, world size=2)
✔️ Relative speed: 1.64 (= 34.1ms / 20.9ms)

[github-actions]

View latest API docs preview at: https://staging.oneflow.info/docs/Oneflow-Inc/oneflow/pr/10325/

[github-actions]

Speed stats:

GPU Name: NVIDIA GeForce RTX 3080 Ti 

❌ OneFlow resnet50 time: 43.7ms (= 4367.6ms / 100, input_shape=[16, 3, 224, 224])
PyTorch resnet50 time: 56.9ms (= 5690.9ms / 100, input_shape=[16, 3, 224, 224])
✔️ Relative speed: 1.30 (= 56.9ms / 43.7ms)

OneFlow resnet50 time: 26.0ms (= 2595.2ms / 100, input_shape=[8, 3, 224, 224])
PyTorch resnet50 time: 39.1ms (= 3905.4ms / 100, input_shape=[8, 3, 224, 224])
✔️ Relative speed: 1.50 (= 39.1ms / 26.0ms)

OneFlow resnet50 time: 18.5ms (= 3692.6ms / 200, input_shape=[4, 3, 224, 224])
PyTorch resnet50 time: 35.8ms (= 7154.9ms / 200, input_shape=[4, 3, 224, 224])
✔️ Relative speed: 1.94 (= 35.8ms / 18.5ms)

OneFlow resnet50 time: 19.0ms (= 3803.9ms / 200, input_shape=[2, 3, 224, 224])
PyTorch resnet50 time: 30.7ms (= 6146.8ms / 200, input_shape=[2, 3, 224, 224])
✔️ Relative speed: 1.62 (= 30.7ms / 19.0ms)

OneFlow resnet50 time: 17.5ms (= 3499.3ms / 200, input_shape=[1, 3, 224, 224])
PyTorch resnet50 time: 29.3ms (= 5861.8ms / 200, input_shape=[1, 3, 224, 224])
✔️ Relative speed: 1.68 (= 29.3ms / 17.5ms)

OneFlow swin dataloader time: 0.203s (= 40.513s / 200, num_workers=1)
PyTorch swin dataloader time: 0.127s (= 25.477s / 200, num_workers=1)
Relative speed: 0.629 (= 0.127s / 0.203s)

OneFlow swin dataloader time: 0.057s (= 11.359s / 200, num_workers=4)
PyTorch swin dataloader time: 0.033s (= 6.556s / 200, num_workers=4)
Relative speed: 0.577 (= 0.033s / 0.057s)

OneFlow swin dataloader time: 0.031s (= 6.278s / 200, num_workers=8)
PyTorch swin dataloader time: 0.017s (= 3.370s / 200, num_workers=8)
Relative speed: 0.537 (= 0.017s / 0.031s)

❌ OneFlow resnet50 time: 47.4ms (= 4744.0ms / 100, input_shape=[16, 3, 224, 224], ddp, world size=2)
PyTorch resnet50 time: 63.0ms (= 6297.7ms / 100, input_shape=[16, 3, 224, 224], ddp, world size=2)
✔️ Relative speed: 1.33 (= 63.0ms / 47.4ms)

OneFlow resnet50 time: 32.5ms (= 3247.7ms / 100, input_shape=[8, 3, 224, 224], ddp, world size=2)
PyTorch resnet50 time: 45.9ms (= 4585.6ms / 100, input_shape=[8, 3, 224, 224], ddp, world size=2)
✔️ Relative speed: 1.41 (= 45.9ms / 32.5ms)

OneFlow resnet50 time: 24.1ms (= 4829.2ms / 200, input_shape=[4, 3, 224, 224], ddp, world size=2)
PyTorch resnet50 time: 42.3ms (= 8458.5ms / 200, input_shape=[4, 3, 224, 224], ddp, world size=2)
✔️ Relative speed: 1.75 (= 42.3ms / 24.1ms)

OneFlow resnet50 time: 22.2ms (= 4437.8ms / 200, input_shape=[2, 3, 224, 224], ddp, world size=2)
PyTorch resnet50 time: 36.9ms (= 7380.8ms / 200, input_shape=[2, 3, 224, 224], ddp, world size=2)
✔️ Relative speed: 1.66 (= 36.9ms / 22.2ms)

OneFlow resnet50 time: 20.6ms (= 4121.9ms / 200, input_shape=[1, 3, 224, 224], ddp, world size=2)
PyTorch resnet50 time: 34.7ms (= 6938.7ms / 200, input_shape=[1, 3, 224, 224], ddp, world size=2)
✔️ Relative speed: 1.68 (= 34.7ms / 20.6ms)