oneflow
oneflow copied to clipboard
Published
20 hours ago •
Oneflow-Inc
Oneflow-Inc
support ci tag: pr sym link
![github-actions[bot] avatar](https://avatars.githubusercontent.com/in/15368?v=4 "Published by a pub.dev verified publisher"){kind=small}
View latest API docs preview at: https://staging.oneflow.info/docs/Oneflow-Inc/oneflow/pr/9119/
Speed stats:
GPU Name: GeForce GTX 1080
❌ OneFlow resnet50 time: 129.6ms (= 12962.6ms / 100, input_shape=[16, 3, 224, 224])
PyTorch resnet50 time: 141.6ms (= 14158.8ms / 100, input_shape=[16, 3, 224, 224])
✔️ Relative speed: 1.09 (= 141.6ms / 129.6ms)
OneFlow resnet50 time: 74.8ms (= 7475.8ms / 100, input_shape=[8, 3, 224, 224])
PyTorch resnet50 time: 85.4ms (= 8542.9ms / 100, input_shape=[8, 3, 224, 224])
✔️ Relative speed: 1.14 (= 85.4ms / 74.8ms)
OneFlow resnet50 time: 47.2ms (= 9445.6ms / 200, input_shape=[4, 3, 224, 224])
PyTorch resnet50 time: 62.5ms (= 12501.4ms / 200, input_shape=[4, 3, 224, 224])
✔️ Relative speed: 1.32 (= 62.5ms / 47.2ms)
OneFlow resnet50 time: 34.5ms (= 6897.4ms / 200, input_shape=[2, 3, 224, 224])
PyTorch resnet50 time: 44.9ms (= 8986.5ms / 200, input_shape=[2, 3, 224, 224])
✔️ Relative speed: 1.30 (= 44.9ms / 34.5ms)
OneFlow resnet50 time: 30.5ms (= 6108.3ms / 200, input_shape=[1, 3, 224, 224])
PyTorch resnet50 time: 38.9ms (= 7779.5ms / 200, input_shape=[1, 3, 224, 224])
✔️ Relative speed: 1.27 (= 38.9ms / 30.5ms)
OneFlow swin dataloader time: 0.256s (= 51.147s / 200, num_workers=1)
PyTorch swin dataloader time: 0.151s (= 30.182s / 200, num_workers=1)
Relative speed: 0.590 (= 0.151s / 0.256s)
OneFlow swin dataloader time: 0.073s (= 14.690s / 200, num_workers=4)
PyTorch swin dataloader time: 0.043s (= 8.532s / 200, num_workers=4)
Relative speed: 0.581 (= 0.043s / 0.073s)
OneFlow swin dataloader time: 0.040s (= 8.007s / 200, num_workers=8)
PyTorch swin dataloader time: 0.022s (= 4.463s / 200, num_workers=8)
Relative speed: 0.557 (= 0.022s / 0.040s)
❌ OneFlow resnet50 time: 139.8ms (= 13983.4ms / 100, input_shape=[16, 3, 224, 224], ddp, world size=2)
PyTorch resnet50 time: 161.0ms (= 16100.5ms / 100, input_shape=[16, 3, 224, 224], ddp, world size=2)
✔️ Relative speed: 1.15 (= 161.0ms / 139.8ms)
OneFlow resnet50 time: 86.0ms (= 8603.5ms / 100, input_shape=[8, 3, 224, 224], ddp, world size=2)
PyTorch resnet50 time: 109.5ms (= 10952.7ms / 100, input_shape=[8, 3, 224, 224], ddp, world size=2)
✔️ Relative speed: 1.27 (= 109.5ms / 86.0ms)
OneFlow resnet50 time: 58.4ms (= 11688.9ms / 200, input_shape=[4, 3, 224, 224], ddp, world size=2)
PyTorch resnet50 time: 78.5ms (= 15695.8ms / 200, input_shape=[4, 3, 224, 224], ddp, world size=2)
✔️ Relative speed: 1.34 (= 78.5ms / 58.4ms)
OneFlow resnet50 time: 45.1ms (= 9020.0ms / 200, input_shape=[2, 3, 224, 224], ddp, world size=2)
PyTorch resnet50 time: 72.3ms (= 14468.0ms / 200, input_shape=[2, 3, 224, 224], ddp, world size=2)
✔️ Relative speed: 1.60 (= 72.3ms / 45.1ms)
OneFlow resnet50 time: 41.3ms (= 8263.8ms / 200, input_shape=[1, 3, 224, 224], ddp, world size=2)
PyTorch resnet50 time: 68.7ms (= 13747.9ms / 200, input_shape=[1, 3, 224, 224], ddp, world size=2)
✔️ Relative speed: 1.66 (= 68.7ms / 41.3ms)
Speed stats:
GPU Name: GeForce GTX 1080
❌ OneFlow resnet50 time: 129.7ms (= 12968.9ms / 100, input_shape=[16, 3, 224, 224])
PyTorch resnet50 time: 142.0ms (= 14199.2ms / 100, input_shape=[16, 3, 224, 224])
✔️ Relative speed: 1.09 (= 142.0ms / 129.7ms)
OneFlow resnet50 time: 74.8ms (= 7478.8ms / 100, input_shape=[8, 3, 224, 224])
PyTorch resnet50 time: 84.6ms (= 8456.7ms / 100, input_shape=[8, 3, 224, 224])
✔️ Relative speed: 1.13 (= 84.6ms / 74.8ms)
OneFlow resnet50 time: 47.1ms (= 9425.4ms / 200, input_shape=[4, 3, 224, 224])
PyTorch resnet50 time: 59.3ms (= 11867.9ms / 200, input_shape=[4, 3, 224, 224])
✔️ Relative speed: 1.26 (= 59.3ms / 47.1ms)
OneFlow resnet50 time: 35.5ms (= 7095.6ms / 200, input_shape=[2, 3, 224, 224])
PyTorch resnet50 time: 42.4ms (= 8483.0ms / 200, input_shape=[2, 3, 224, 224])
✔️ Relative speed: 1.20 (= 42.4ms / 35.5ms)
OneFlow resnet50 time: 29.7ms (= 5933.6ms / 200, input_shape=[1, 3, 224, 224])
PyTorch resnet50 time: 37.1ms (= 7416.9ms / 200, input_shape=[1, 3, 224, 224])
✔️ Relative speed: 1.25 (= 37.1ms / 29.7ms)
OneFlow swin dataloader time: 0.256s (= 51.120s / 200, num_workers=1)
PyTorch swin dataloader time: 0.147s (= 29.430s / 200, num_workers=1)
Relative speed: 0.576 (= 0.147s / 0.256s)
OneFlow swin dataloader time: 0.072s (= 14.354s / 200, num_workers=4)
PyTorch swin dataloader time: 0.040s (= 7.976s / 200, num_workers=4)
Relative speed: 0.556 (= 0.040s / 0.072s)
OneFlow swin dataloader time: 0.040s (= 7.911s / 200, num_workers=8)
PyTorch swin dataloader time: 0.022s (= 4.368s / 200, num_workers=8)
Relative speed: 0.552 (= 0.022s / 0.040s)
❌ OneFlow resnet50 time: 140.1ms (= 14007.2ms / 100, input_shape=[16, 3, 224, 224], ddp, world size=2)
PyTorch resnet50 time: 161.7ms (= 16170.4ms / 100, input_shape=[16, 3, 224, 224], ddp, world size=2)
✔️ Relative speed: 1.15 (= 161.7ms / 140.1ms)
OneFlow resnet50 time: 85.6ms (= 8559.9ms / 100, input_shape=[8, 3, 224, 224], ddp, world size=2)
PyTorch resnet50 time: 102.3ms (= 10226.5ms / 100, input_shape=[8, 3, 224, 224], ddp, world size=2)
✔️ Relative speed: 1.19 (= 102.3ms / 85.6ms)
OneFlow resnet50 time: 58.3ms (= 11666.6ms / 200, input_shape=[4, 3, 224, 224], ddp, world size=2)
PyTorch resnet50 time: 78.5ms (= 15698.5ms / 200, input_shape=[4, 3, 224, 224], ddp, world size=2)
✔️ Relative speed: 1.35 (= 78.5ms / 58.3ms)
OneFlow resnet50 time: 46.1ms (= 9216.2ms / 200, input_shape=[2, 3, 224, 224], ddp, world size=2)
PyTorch resnet50 time: 72.8ms (= 14569.2ms / 200, input_shape=[2, 3, 224, 224], ddp, world size=2)
✔️ Relative speed: 1.58 (= 72.8ms / 46.1ms)
OneFlow resnet50 time: 41.3ms (= 8260.0ms / 200, input_shape=[1, 3, 224, 224], ddp, world size=2)
PyTorch resnet50 time: 63.5ms (= 12703.1ms / 200, input_shape=[1, 3, 224, 224], ddp, world size=2)
✔️ Relative speed: 1.54 (= 63.5ms / 41.3ms)
View latest API docs preview at: https://staging.oneflow.info/docs/Oneflow-Inc/oneflow/pr/9119/
View latest API docs preview at: https://staging.oneflow.info/docs/Oneflow-Inc/oneflow/pr/9119/
Speed stats:
GPU Name: GeForce GTX 1080
❌ OneFlow resnet50 time: 129.8ms (= 12981.6ms / 100, input_shape=[16, 3, 224, 224])
PyTorch resnet50 time: 142.7ms (= 14269.7ms / 100, input_shape=[16, 3, 224, 224])
✔️ Relative speed: 1.10 (= 142.7ms / 129.8ms)
OneFlow resnet50 time: 74.9ms (= 7490.6ms / 100, input_shape=[8, 3, 224, 224])
PyTorch resnet50 time: 86.5ms (= 8652.3ms / 100, input_shape=[8, 3, 224, 224])
✔️ Relative speed: 1.16 (= 86.5ms / 74.9ms)
OneFlow resnet50 time: 47.8ms (= 9554.4ms / 200, input_shape=[4, 3, 224, 224])
PyTorch resnet50 time: 60.7ms (= 12145.5ms / 200, input_shape=[4, 3, 224, 224])
✔️ Relative speed: 1.27 (= 60.7ms / 47.8ms)
OneFlow resnet50 time: 34.9ms (= 6987.3ms / 200, input_shape=[2, 3, 224, 224])
PyTorch resnet50 time: 45.4ms (= 9083.4ms / 200, input_shape=[2, 3, 224, 224])
✔️ Relative speed: 1.30 (= 45.4ms / 34.9ms)
OneFlow resnet50 time: 29.9ms (= 5970.1ms / 200, input_shape=[1, 3, 224, 224])
PyTorch resnet50 time: 36.7ms (= 7342.9ms / 200, input_shape=[1, 3, 224, 224])
✔️ Relative speed: 1.23 (= 36.7ms / 29.9ms)
OneFlow swin dataloader time: 0.261s (= 52.228s / 200, num_workers=1)
PyTorch swin dataloader time: 0.149s (= 29.827s / 200, num_workers=1)
Relative speed: 0.571 (= 0.149s / 0.261s)
OneFlow swin dataloader time: 0.074s (= 14.716s / 200, num_workers=4)
PyTorch swin dataloader time: 0.041s (= 8.225s / 200, num_workers=4)
Relative speed: 0.559 (= 0.041s / 0.074s)
OneFlow swin dataloader time: 0.043s (= 8.588s / 200, num_workers=8)
PyTorch swin dataloader time: 0.022s (= 4.422s / 200, num_workers=8)
Relative speed: 0.515 (= 0.022s / 0.043s)
❌ OneFlow resnet50 time: 140.0ms (= 13998.3ms / 100, input_shape=[16, 3, 224, 224], ddp, world size=2)
PyTorch resnet50 time: 162.1ms (= 16205.6ms / 100, input_shape=[16, 3, 224, 224], ddp, world size=2)
✔️ Relative speed: 1.16 (= 162.1ms / 140.0ms)
OneFlow resnet50 time: 86.1ms (= 8608.6ms / 100, input_shape=[8, 3, 224, 224], ddp, world size=2)
PyTorch resnet50 time: 102.4ms (= 10237.3ms / 100, input_shape=[8, 3, 224, 224], ddp, world size=2)
✔️ Relative speed: 1.19 (= 102.4ms / 86.1ms)
OneFlow resnet50 time: 58.4ms (= 11680.9ms / 200, input_shape=[4, 3, 224, 224], ddp, world size=2)
PyTorch resnet50 time: 78.1ms (= 15616.3ms / 200, input_shape=[4, 3, 224, 224], ddp, world size=2)
✔️ Relative speed: 1.34 (= 78.1ms / 58.4ms)
OneFlow resnet50 time: 44.7ms (= 8938.2ms / 200, input_shape=[2, 3, 224, 224], ddp, world size=2)
PyTorch resnet50 time: 78.4ms (= 15686.6ms / 200, input_shape=[2, 3, 224, 224], ddp, world size=2)
✔️ Relative speed: 1.76 (= 78.4ms / 44.7ms)
OneFlow resnet50 time: 40.5ms (= 8092.2ms / 200, input_shape=[1, 3, 224, 224], ddp, world size=2)
PyTorch resnet50 time: 75.3ms (= 15068.0ms / 200, input_shape=[1, 3, 224, 224], ddp, world size=2)
✔️ Relative speed: 1.86 (= 75.3ms / 40.5ms)
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.
View latest API docs preview at: https://staging.oneflow.info/docs/Oneflow-Inc/oneflow/pr/9119/
Speed stats:
GPU Name: GeForce GTX 1080
❌ OneFlow resnet50 time: 140.2ms (= 14015.8ms / 100, input_shape=[16, 3, 224, 224], ddp, world size=2)
PyTorch resnet50 time: 167.2ms (= 16718.9ms / 100, input_shape=[16, 3, 224, 224], ddp, world size=2)
✔️ Relative speed: 1.19 (= 167.2ms / 140.2ms)
OneFlow resnet50 time: 85.6ms (= 8560.0ms / 100, input_shape=[8, 3, 224, 224], ddp, world size=2)
PyTorch resnet50 time: 110.7ms (= 11068.0ms / 100, input_shape=[8, 3, 224, 224], ddp, world size=2)
✔️ Relative speed: 1.29 (= 110.7ms / 85.6ms)
OneFlow resnet50 time: 57.8ms (= 11559.2ms / 200, input_shape=[4, 3, 224, 224], ddp, world size=2)
PyTorch resnet50 time: 78.4ms (= 15681.0ms / 200, input_shape=[4, 3, 224, 224], ddp, world size=2)
✔️ Relative speed: 1.36 (= 78.4ms / 57.8ms)
OneFlow resnet50 time: 44.9ms (= 8985.9ms / 200, input_shape=[2, 3, 224, 224], ddp, world size=2)
PyTorch resnet50 time: 72.7ms (= 14535.0ms / 200, input_shape=[2, 3, 224, 224], ddp, world size=2)
✔️ Relative speed: 1.62 (= 72.7ms / 44.9ms)
OneFlow resnet50 time: 39.3ms (= 7853.5ms / 200, input_shape=[1, 3, 224, 224], ddp, world size=2)
PyTorch resnet50 time: 68.5ms (= 13690.9ms / 200, input_shape=[1, 3, 224, 224], ddp, world size=2)
✔️ Relative speed: 1.74 (= 68.5ms / 39.3ms)
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.
View latest API docs preview at: https://staging.oneflow.info/docs/Oneflow-Inc/oneflow/pr/9119/
Speed stats:
GPU Name: GeForce GTX 1080
❌ OneFlow resnet50 time: 139.8ms (= 13980.4ms / 100, input_shape=[16, 3, 224, 224], ddp, world size=2)
PyTorch resnet50 time: 162.4ms (= 16239.1ms / 100, input_shape=[16, 3, 224, 224], ddp, world size=2)
✔️ Relative speed: 1.16 (= 162.4ms / 139.8ms)
OneFlow resnet50 time: 85.6ms (= 8564.8ms / 100, input_shape=[8, 3, 224, 224], ddp, world size=2)
PyTorch resnet50 time: 101.4ms (= 10142.4ms / 100, input_shape=[8, 3, 224, 224], ddp, world size=2)
✔️ Relative speed: 1.18 (= 101.4ms / 85.6ms)
OneFlow resnet50 time: 58.4ms (= 11678.2ms / 200, input_shape=[4, 3, 224, 224], ddp, world size=2)
PyTorch resnet50 time: 78.8ms (= 15764.9ms / 200, input_shape=[4, 3, 224, 224], ddp, world size=2)
✔️ Relative speed: 1.35 (= 78.8ms / 58.4ms)
OneFlow resnet50 time: 44.8ms (= 8969.5ms / 200, input_shape=[2, 3, 224, 224], ddp, world size=2)
PyTorch resnet50 time: 69.9ms (= 13972.3ms / 200, input_shape=[2, 3, 224, 224], ddp, world size=2)
✔️ Relative speed: 1.56 (= 69.9ms / 44.8ms)
OneFlow resnet50 time: 39.7ms (= 7943.0ms / 200, input_shape=[1, 3, 224, 224], ddp, world size=2)
PyTorch resnet50 time: 68.5ms (= 13709.2ms / 200, input_shape=[1, 3, 224, 224], ddp, world size=2)
✔️ Relative speed: 1.73 (= 68.5ms / 39.7ms)
Nginx seems to fail handling symbolic link even the disymlink is off on ci machine, maybe caused by the mount file system.