Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom

dev_median #8069

Merged
merged 20 commits into from
Apr 27, 2022
Merged

dev_median #8069

merged 20 commits into from
Apr 27, 2022

Conversation

simonJJJ
Copy link
Contributor

No description provided.

@simonJJJ simonJJJ marked this pull request as ready for review April 21, 2022 08:38
BBuf
BBuf reviewed Apr 22, 2022
View reviewed changes
@simonJJJ simonJJJ requested a review from oneflow-ci-bot April 25, 2022 09:48
@oneflow-ci-bot oneflow-ci-bot requested review from oneflow-ci-bot and removed request for oneflow-ci-bot April 25, 2022 17:22
@github-actions
Copy link
Contributor

View latest API docs preview at: https://staging.oneflow.info/docs/Oneflow-Inc/oneflow/pr/8069/

@github-actions
Copy link
Contributor

Speed stats: 
GPU Name: GeForce GTX 1080 

✔️ OneFlow resnet50 time: 128.4ms (= 12843.2ms / 100, input_shape=[16, 3, 224, 224])
PyTorch resnet50 time: 141.4ms (= 14138.5ms / 100, input_shape=[16, 3, 224, 224])
✔️ Relative speed: 1.10 (= 141.4ms / 128.4ms)

OneFlow resnet50 time: 79.0ms (= 7901.4ms / 100, input_shape=[8, 3, 224, 224])
PyTorch resnet50 time: 85.0ms (= 8504.3ms / 100, input_shape=[8, 3, 224, 224])
✔️ Relative speed: 1.08 (= 85.0ms / 79.0ms)

OneFlow resnet50 time: 54.5ms (= 10904.7ms / 200, input_shape=[4, 3, 224, 224])
PyTorch resnet50 time: 61.9ms (= 12373.5ms / 200, input_shape=[4, 3, 224, 224])
✔️ Relative speed: 1.13 (= 61.9ms / 54.5ms)

OneFlow resnet50 time: 41.8ms (= 8359.5ms / 200, input_shape=[2, 3, 224, 224])
PyTorch resnet50 time: 45.6ms (= 9123.7ms / 200, input_shape=[2, 3, 224, 224])
✔️ Relative speed: 1.09 (= 45.6ms / 41.8ms)

OneFlow resnet50 time: 34.6ms (= 6921.0ms / 200, input_shape=[1, 3, 224, 224])
PyTorch resnet50 time: 36.0ms (= 7192.4ms / 200, input_shape=[1, 3, 224, 224])
✔️ Relative speed: 1.04 (= 36.0ms / 34.6ms)

OneFlow swin dataloader time: 0.246s (= 49.255s / 200, num_workers=1)
PyTorch swin dataloader time: 0.150s (= 29.907s / 200, num_workers=1)
Relative speed: 0.607 (= 0.150s / 0.246s)

OneFlow swin dataloader time: 0.069s (= 13.743s / 200, num_workers=4)
PyTorch swin dataloader time: 0.041s (= 8.168s / 200, num_workers=4)
Relative speed: 0.594 (= 0.041s / 0.069s)

OneFlow swin dataloader time: 0.036s (= 7.215s / 200, num_workers=8)
PyTorch swin dataloader time: 0.022s (= 4.393s / 200, num_workers=8)
Relative speed: 0.609 (= 0.022s / 0.036s)

✔️ OneFlow resnet50 time: 135.9ms (= 13593.5ms / 100, input_shape=[16, 3, 224, 224], ddp, world size=2)
PyTorch resnet50 time: 162.2ms (= 16215.6ms / 100, input_shape=[16, 3, 224, 224], ddp, world size=2)
✔️ Relative speed: 1.19 (= 162.2ms / 135.9ms)

OneFlow resnet50 time: 88.6ms (= 8862.9ms / 100, input_shape=[8, 3, 224, 224], ddp, world size=2)
PyTorch resnet50 time: 100.3ms (= 10032.6ms / 100, input_shape=[8, 3, 224, 224], ddp, world size=2)
✔️ Relative speed: 1.13 (= 100.3ms / 88.6ms)

OneFlow resnet50 time: 64.4ms (= 12873.6ms / 200, input_shape=[4, 3, 224, 224], ddp, world size=2)
PyTorch resnet50 time: 78.4ms (= 15672.4ms / 200, input_shape=[4, 3, 224, 224], ddp, world size=2)
✔️ Relative speed: 1.22 (= 78.4ms / 64.4ms)

OneFlow resnet50 time: 50.3ms (= 10055.7ms / 200, input_shape=[2, 3, 224, 224], ddp, world size=2)
PyTorch resnet50 time: 78.6ms (= 15713.9ms / 200, input_shape=[2, 3, 224, 224], ddp, world size=2)
✔️ Relative speed: 1.56 (= 78.6ms / 50.3ms)

OneFlow resnet50 time: 46.4ms (= 9274.7ms / 200, input_shape=[1, 3, 224, 224], ddp, world size=2)
PyTorch resnet50 time: 66.5ms (= 13293.4ms / 200, input_shape=[1, 3, 224, 224], ddp, world size=2)
✔️ Relative speed: 1.43 (= 66.5ms / 46.4ms)

@github-actions
Copy link
Contributor

View latest API docs preview at: https://staging.oneflow.info/docs/Oneflow-Inc/oneflow/pr/8069/

@github-actions
Copy link
Contributor

Speed stats: 
GPU Name: GeForce GTX 1080 

✔️ OneFlow resnet50 time: 128.3ms (= 12833.6ms / 100, input_shape=[16, 3, 224, 224])
PyTorch resnet50 time: 141.0ms (= 14102.5ms / 100, input_shape=[16, 3, 224, 224])
✔️ Relative speed: 1.10 (= 141.0ms / 128.3ms)

OneFlow resnet50 time: 77.6ms (= 7761.6ms / 100, input_shape=[8, 3, 224, 224])
PyTorch resnet50 time: 85.3ms (= 8534.4ms / 100, input_shape=[8, 3, 224, 224])
✔️ Relative speed: 1.10 (= 85.3ms / 77.6ms)

OneFlow resnet50 time: 51.9ms (= 10378.6ms / 200, input_shape=[4, 3, 224, 224])
PyTorch resnet50 time: 62.6ms (= 12524.4ms / 200, input_shape=[4, 3, 224, 224])
✔️ Relative speed: 1.21 (= 62.6ms / 51.9ms)

OneFlow resnet50 time: 39.8ms (= 7952.1ms / 200, input_shape=[2, 3, 224, 224])
PyTorch resnet50 time: 40.5ms (= 8104.7ms / 200, input_shape=[2, 3, 224, 224])
✔️ Relative speed: 1.02 (= 40.5ms / 39.8ms)

OneFlow resnet50 time: 37.9ms (= 7570.0ms / 200, input_shape=[1, 3, 224, 224])
PyTorch resnet50 time: 36.2ms (= 7247.1ms / 200, input_shape=[1, 3, 224, 224])
✔️ Relative speed: 0.96 (= 36.2ms / 37.9ms)

OneFlow swin dataloader time: 0.257s (= 51.462s / 200, num_workers=1)
PyTorch swin dataloader time: 0.149s (= 29.844s / 200, num_workers=1)
Relative speed: 0.580 (= 0.149s / 0.257s)

OneFlow swin dataloader time: 0.066s (= 13.222s / 200, num_workers=4)
PyTorch swin dataloader time: 0.041s (= 8.107s / 200, num_workers=4)
Relative speed: 0.613 (= 0.041s / 0.066s)

OneFlow swin dataloader time: 0.038s (= 7.566s / 200, num_workers=8)
PyTorch swin dataloader time: 0.023s (= 4.663s / 200, num_workers=8)
Relative speed: 0.616 (= 0.023s / 0.038s)

✔️ OneFlow resnet50 time: 135.7ms (= 13565.1ms / 100, input_shape=[16, 3, 224, 224], ddp, world size=2)
PyTorch resnet50 time: 167.6ms (= 16755.2ms / 100, input_shape=[16, 3, 224, 224], ddp, world size=2)
✔️ Relative speed: 1.24 (= 167.6ms / 135.7ms)

OneFlow resnet50 time: 83.4ms (= 8338.8ms / 100, input_shape=[8, 3, 224, 224], ddp, world size=2)
PyTorch resnet50 time: 112.2ms (= 11215.9ms / 100, input_shape=[8, 3, 224, 224], ddp, world size=2)
✔️ Relative speed: 1.35 (= 112.2ms / 83.4ms)

OneFlow resnet50 time: 63.5ms (= 12706.4ms / 200, input_shape=[4, 3, 224, 224], ddp, world size=2)
PyTorch resnet50 time: 91.3ms (= 18262.3ms / 200, input_shape=[4, 3, 224, 224], ddp, world size=2)
✔️ Relative speed: 1.44 (= 91.3ms / 63.5ms)

OneFlow resnet50 time: 52.5ms (= 10505.8ms / 200, input_shape=[2, 3, 224, 224], ddp, world size=2)
PyTorch resnet50 time: 78.2ms (= 15641.3ms / 200, input_shape=[2, 3, 224, 224], ddp, world size=2)
✔️ Relative speed: 1.49 (= 78.2ms / 52.5ms)

OneFlow resnet50 time: 48.0ms (= 9597.6ms / 200, input_shape=[1, 3, 224, 224], ddp, world size=2)
PyTorch resnet50 time: 66.4ms (= 13282.2ms / 200, input_shape=[1, 3, 224, 224], ddp, world size=2)
✔️ Relative speed: 1.38 (= 66.4ms / 48.0ms)

@github-actions
Copy link
Contributor

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

@github-actions
Copy link
Contributor

CI failed when running job: cpu-module. PR label automerge has been removed

@github-actions
Copy link
Contributor

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
Copy link
Contributor

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
Copy link
Contributor

View latest API docs preview at: https://staging.oneflow.info/docs/Oneflow-Inc/oneflow/pr/8069/

@github-actions
Copy link
Contributor

Speed stats: 
GPU Name: NVIDIA GeForce GTX 1080 

❌ OneFlow resnet50 time: 129.3ms (= 12933.9ms / 100, input_shape=[16, 3, 224, 224])
PyTorch resnet50 time: 144.3ms (= 14426.8ms / 100, input_shape=[16, 3, 224, 224])
✔️ Relative speed: 1.12 (= 144.3ms / 129.3ms)

OneFlow resnet50 time: 78.1ms (= 7814.4ms / 100, input_shape=[8, 3, 224, 224])
PyTorch resnet50 time: 84.6ms (= 8456.2ms / 100, input_shape=[8, 3, 224, 224])
✔️ Relative speed: 1.08 (= 84.6ms / 78.1ms)

OneFlow resnet50 time: 49.1ms (= 9822.9ms / 200, input_shape=[4, 3, 224, 224])
PyTorch resnet50 time: 61.2ms (= 12247.0ms / 200, input_shape=[4, 3, 224, 224])
✔️ Relative speed: 1.25 (= 61.2ms / 49.1ms)

OneFlow resnet50 time: 42.8ms (= 8553.5ms / 200, input_shape=[2, 3, 224, 224])
PyTorch resnet50 time: 41.2ms (= 8237.9ms / 200, input_shape=[2, 3, 224, 224])
❌ Relative speed: 0.96 (= 41.2ms / 42.8ms)

OneFlow resnet50 time: 38.3ms (= 7661.3ms / 200, input_shape=[1, 3, 224, 224])
PyTorch resnet50 time: 43.8ms (= 8770.0ms / 200, input_shape=[1, 3, 224, 224])
✔️ Relative speed: 1.14 (= 43.8ms / 38.3ms)

OneFlow swin dataloader time: 0.252s (= 50.492s / 200, num_workers=1)
PyTorch swin dataloader time: 0.148s (= 29.594s / 200, num_workers=1)
Relative speed: 0.586 (= 0.148s / 0.252s)

OneFlow swin dataloader time: 0.067s (= 13.461s / 200, num_workers=4)
PyTorch swin dataloader time: 0.041s (= 8.228s / 200, num_workers=4)
Relative speed: 0.611 (= 0.041s / 0.067s)

OneFlow swin dataloader time: 0.037s (= 7.456s / 200, num_workers=8)
PyTorch swin dataloader time: 0.023s (= 4.624s / 200, num_workers=8)
Relative speed: 0.620 (= 0.023s / 0.037s)

❌ OneFlow resnet50 time: 145.1ms (= 14511.0ms / 100, input_shape=[16, 3, 224, 224], ddp, world size=2)
PyTorch resnet50 time: 169.1ms (= 16908.2ms / 100, input_shape=[16, 3, 224, 224], ddp, world size=2)
✔️ Relative speed: 1.17 (= 169.1ms / 145.1ms)

OneFlow resnet50 time: 93.0ms (= 9304.1ms / 100, input_shape=[8, 3, 224, 224], ddp, world size=2)
PyTorch resnet50 time: 111.2ms (= 11119.8ms / 100, input_shape=[8, 3, 224, 224], ddp, world size=2)
✔️ Relative speed: 1.20 (= 111.2ms / 93.0ms)

OneFlow resnet50 time: 73.7ms (= 14739.4ms / 200, input_shape=[4, 3, 224, 224], ddp, world size=2)
PyTorch resnet50 time: 88.7ms (= 17730.7ms / 200, input_shape=[4, 3, 224, 224], ddp, world size=2)
✔️ Relative speed: 1.20 (= 88.7ms / 73.7ms)

OneFlow resnet50 time: 64.4ms (= 12877.7ms / 200, input_shape=[2, 3, 224, 224], ddp, world size=2)
PyTorch resnet50 time: 74.7ms (= 14931.2ms / 200, input_shape=[2, 3, 224, 224], ddp, world size=2)
❌ Relative speed: 1.16 (= 74.7ms / 64.4ms)

OneFlow resnet50 time: 56.4ms (= 11286.8ms / 200, input_shape=[1, 3, 224, 224], ddp, world size=2)
PyTorch resnet50 time: 74.4ms (= 14882.4ms / 200, input_shape=[1, 3, 224, 224], ddp, world size=2)
✔️ Relative speed: 1.32 (= 74.4ms / 56.4ms)

@mergify mergify bot merged commit 53fb647 into master Apr 27, 2022
@mergify mergify bot deleted the dev_median branch April 27, 2022 21:43
@zhongshsh zhongshsh mentioned this pull request May 6, 2022
Closed
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Reviewers

@Flowingsun007 Flowingsun007 Flowingsun007 approved these changes

@BBuf BBuf BBuf approved these changes

@doombeaker doombeaker Awaiting requested review from doombeaker doombeaker is a code owner

@jackalcooper jackalcooper Awaiting requested review from jackalcooper jackalcooper is a code owner

@hjchen2 hjchen2 Awaiting requested review from hjchen2 hjchen2 is a code owner

@daquexian daquexian Awaiting requested review from daquexian

@liujuncheng liujuncheng Awaiting requested review from liujuncheng liujuncheng is a code owner

@guo-ran guo-ran Awaiting requested review from guo-ran

@MARD1NO MARD1NO Awaiting requested review from MARD1NO

@oneflow-ci-bot oneflow-ci-bot Awaiting requested review from oneflow-ci-bot

Assignees
No one assigned
Labels
api automerge eager feature op
Projects
None yet
Milestone
No milestone
Development

Successfully merging this pull request may close these issues.
4 participants
@simonJJJ @Flowingsun007 @BBuf @oneflow-ci-bot