Remove legacy model io

[liujuncheng]

No description provided.

[github-actions]

Code got formatted by CI. Please request CI again if you still want to have this PR merged. If the PR is from a forked repo, please download the patch files from the GitHub Actions web page and apply them locally.

[github-actions]

View latest API docs preview at: https://staging.oneflow.info/docs/Oneflow-Inc/oneflow/pr/8151/

[github-actions]

Speed stats:

GPU Name: NVIDIA GeForce GTX 1080 

❌ OneFlow resnet50 time: 129.8ms (= 12984.5ms / 100, input_shape=[16, 3, 224, 224])
PyTorch resnet50 time: 143.6ms (= 14355.2ms / 100, input_shape=[16, 3, 224, 224])
✔️ Relative speed: 1.11 (= 143.6ms / 129.8ms)

OneFlow resnet50 time: 80.7ms (= 8071.6ms / 100, input_shape=[8, 3, 224, 224])
PyTorch resnet50 time: 83.9ms (= 8386.8ms / 100, input_shape=[8, 3, 224, 224])
✔️ Relative speed: 1.04 (= 83.9ms / 80.7ms)

OneFlow resnet50 time: 52.0ms (= 10400.4ms / 200, input_shape=[4, 3, 224, 224])
PyTorch resnet50 time: 61.6ms (= 12311.3ms / 200, input_shape=[4, 3, 224, 224])
✔️ Relative speed: 1.18 (= 61.6ms / 52.0ms)

OneFlow resnet50 time: 43.8ms (= 8769.7ms / 200, input_shape=[2, 3, 224, 224])
PyTorch resnet50 time: 42.1ms (= 8415.0ms / 200, input_shape=[2, 3, 224, 224])
❌ Relative speed: 0.96 (= 42.1ms / 43.8ms)

OneFlow resnet50 time: 36.1ms (= 7211.3ms / 200, input_shape=[1, 3, 224, 224])
PyTorch resnet50 time: 37.2ms (= 7449.0ms / 200, input_shape=[1, 3, 224, 224])
✔️ Relative speed: 1.03 (= 37.2ms / 36.1ms)

OneFlow swin dataloader time: 0.256s (= 51.259s / 200, num_workers=1)
PyTorch swin dataloader time: 0.151s (= 30.135s / 200, num_workers=1)
Relative speed: 0.588 (= 0.151s / 0.256s)

OneFlow swin dataloader time: 0.067s (= 13.313s / 200, num_workers=4)
PyTorch swin dataloader time: 0.042s (= 8.385s / 200, num_workers=4)
Relative speed: 0.630 (= 0.042s / 0.067s)

OneFlow swin dataloader time: 0.039s (= 7.728s / 200, num_workers=8)
PyTorch swin dataloader time: 0.023s (= 4.512s / 200, num_workers=8)
Relative speed: 0.584 (= 0.023s / 0.039s)

❌ OneFlow resnet50 time: 145.8ms (= 14578.8ms / 100, input_shape=[16, 3, 224, 224], ddp, world size=2)
PyTorch resnet50 time: 168.9ms (= 16888.9ms / 100, input_shape=[16, 3, 224, 224], ddp, world size=2)
✔️ Relative speed: 1.16 (= 168.9ms / 145.8ms)

OneFlow resnet50 time: 96.2ms (= 9620.4ms / 100, input_shape=[8, 3, 224, 224], ddp, world size=2)
PyTorch resnet50 time: 111.8ms (= 11184.1ms / 100, input_shape=[8, 3, 224, 224], ddp, world size=2)
✔️ Relative speed: 1.16 (= 111.8ms / 96.2ms)

OneFlow resnet50 time: 76.1ms (= 15210.1ms / 200, input_shape=[4, 3, 224, 224], ddp, world size=2)
PyTorch resnet50 time: 87.3ms (= 17464.1ms / 200, input_shape=[4, 3, 224, 224], ddp, world size=2)
❌ Relative speed: 1.15 (= 87.3ms / 76.1ms)

OneFlow resnet50 time: 60.7ms (= 12133.4ms / 200, input_shape=[2, 3, 224, 224], ddp, world size=2)
PyTorch resnet50 time: 79.5ms (= 15896.1ms / 200, input_shape=[2, 3, 224, 224], ddp, world size=2)
✔️ Relative speed: 1.31 (= 79.5ms / 60.7ms)

OneFlow resnet50 time: 57.0ms (= 11393.1ms / 200, input_shape=[1, 3, 224, 224], ddp, world size=2)
PyTorch resnet50 time: 79.0ms (= 15794.1ms / 200, input_shape=[1, 3, 224, 224], ddp, world size=2)
✔️ Relative speed: 1.39 (= 79.0ms / 57.0ms)

[github-actions]

CI failed when running job: cuda-benchmark. PR label automerge has been removed