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

[For dev]Sdxl dev #10323

Closed
wants to merge 16 commits into from
Closed

[For dev]Sdxl dev #10323

wants to merge 16 commits into from

Conversation

strint
Copy link
Contributor

@strint strint commented Aug 25, 2023 •
edited
Loading

strint and others added 15 commits August 23, 2023 04:47
@strint
fake dynamo
30435f9
@strint
Merge branch 'master' of https://github.com/Oneflow-Inc/oneflow into …
6cfa893 
…fix_dynamo_mock_error
@oneflow-ci-bot
auto format by CI
69bbc0e
@strint
general argstree
998c501
@strint
Merge branch 'fix_dynamo_mock_error' of https://github.com/Oneflow-In…
eccb5f1 
…c/oneflow into fix_dynamo_mock_error
@oneflow-ci-bot
auto format by CI
44a3710
@strint
refine repr of argstree
3619406
@strint
Merge branch 'fix_dynamo_mock_error' of https://github.com/Oneflow-In…
93d0fe3 
…c/oneflow into fix_dynamo_mock_error
@oneflow-ci-bot
auto format by CI
ccac35e
@strint
graph allow any input
025d4c5
@strint
Merge branch 'fix_dynamo_mock_error' of https://github.com/Oneflow-In…
679e0d5 
…c/oneflow into fix_dynamo_mock_error
@strint
save support any input
354cf21
@jackalcooper
fix
99b3d60
@jackalcooper
fix gsize>64
8347ed8
@strint
Merge branch 'add-kMaxInputCount-in-GroupedMatmulFunctor' of https://…
6705366 
…github.com/Oneflow-Inc/oneflow into sdxl_dev
@strint strint mentioned this pull request Aug 25, 2023
Merged
@hjchen2 hjchen2 enabled auto-merge (squash) August 25, 2023 12:06
@github-actions
Copy link
Contributor

View latest API docs preview at: https://staging.oneflow.info/docs/Oneflow-Inc/oneflow/pr/10323/

@github-actions
Copy link
Contributor

Speed stats: 
GPU Name: NVIDIA GeForce RTX 3080 Ti 

❌ OneFlow resnet50 time: 43.5ms (= 4350.9ms / 100, input_shape=[16, 3, 224, 224])
PyTorch resnet50 time: 61.4ms (= 6136.9ms / 100, input_shape=[16, 3, 224, 224])
✔️ Relative speed: 1.41 (= 61.4ms / 43.5ms)

OneFlow resnet50 time: 25.9ms (= 2592.6ms / 100, input_shape=[8, 3, 224, 224])
PyTorch resnet50 time: 37.5ms (= 3752.6ms / 100, input_shape=[8, 3, 224, 224])
✔️ Relative speed: 1.45 (= 37.5ms / 25.9ms)

OneFlow resnet50 time: 19.1ms (= 3827.3ms / 200, input_shape=[4, 3, 224, 224])
PyTorch resnet50 time: 34.8ms (= 6967.0ms / 200, input_shape=[4, 3, 224, 224])
✔️ Relative speed: 1.82 (= 34.8ms / 19.1ms)

OneFlow resnet50 time: 19.2ms (= 3846.6ms / 200, input_shape=[2, 3, 224, 224])
PyTorch resnet50 time: 32.5ms (= 6495.0ms / 200, input_shape=[2, 3, 224, 224])
✔️ Relative speed: 1.69 (= 32.5ms / 19.2ms)

OneFlow resnet50 time: 17.6ms (= 3510.1ms / 200, input_shape=[1, 3, 224, 224])
PyTorch resnet50 time: 29.8ms (= 5952.7ms / 200, input_shape=[1, 3, 224, 224])
✔️ Relative speed: 1.70 (= 29.8ms / 17.6ms)

OneFlow swin dataloader time: 0.201s (= 40.143s / 200, num_workers=1)
PyTorch swin dataloader time: 0.127s (= 25.405s / 200, num_workers=1)
Relative speed: 0.633 (= 0.127s / 0.201s)

OneFlow swin dataloader time: 0.055s (= 10.984s / 200, num_workers=4)
PyTorch swin dataloader time: 0.033s (= 6.630s / 200, num_workers=4)
Relative speed: 0.604 (= 0.033s / 0.055s)

OneFlow swin dataloader time: 0.031s (= 6.206s / 200, num_workers=8)
PyTorch swin dataloader time: 0.017s (= 3.474s / 200, num_workers=8)
Relative speed: 0.560 (= 0.017s / 0.031s)

❌ OneFlow resnet50 time: 47.6ms (= 4762.6ms / 100, input_shape=[16, 3, 224, 224], ddp, world size=2)
PyTorch resnet50 time: 63.7ms (= 6367.8ms / 100, input_shape=[16, 3, 224, 224], ddp, world size=2)
✔️ Relative speed: 1.34 (= 63.7ms / 47.6ms)

OneFlow resnet50 time: 31.1ms (= 3108.7ms / 100, input_shape=[8, 3, 224, 224], ddp, world size=2)
PyTorch resnet50 time: 43.5ms (= 4347.0ms / 100, input_shape=[8, 3, 224, 224], ddp, world size=2)
✔️ Relative speed: 1.40 (= 43.5ms / 31.1ms)

OneFlow resnet50 time: 24.2ms (= 4839.6ms / 200, input_shape=[4, 3, 224, 224], ddp, world size=2)
PyTorch resnet50 time: 41.0ms (= 8199.7ms / 200, input_shape=[4, 3, 224, 224], ddp, world size=2)
✔️ Relative speed: 1.69 (= 41.0ms / 24.2ms)

OneFlow resnet50 time: 22.6ms (= 4525.8ms / 200, input_shape=[2, 3, 224, 224], ddp, world size=2)
PyTorch resnet50 time: 38.1ms (= 7621.2ms / 200, input_shape=[2, 3, 224, 224], ddp, world size=2)
✔️ Relative speed: 1.68 (= 38.1ms / 22.6ms)

OneFlow resnet50 time: 20.8ms (= 4150.7ms / 200, input_shape=[1, 3, 224, 224], ddp, world size=2)
PyTorch resnet50 time: 34.3ms (= 6850.1ms / 200, input_shape=[1, 3, 224, 224], ddp, world size=2)
✔️ Relative speed: 1.65 (= 34.3ms / 20.8ms)

hjchen2 pushed a commit to siliconflow/onediff that referenced this pull request Aug 29, 2023
@strint
sd2 oneflow compile (#244)
3680e73 
Depends on 
Oneflow-Inc/oneflow#10323

### unet graph(sd2)
run
```
python examples/unet_torch_interplay.py --model_id=stabilityai/stable-diffusion-xl-base-1.0 --revision=main
```
run and save
```
python examples/unet_torch_interplay.py --model_id=stabilityai/stable-diffusion-xl-base-1.0 --revision=main --save
```
load and run
```
python examples/unet_torch_interplay.py --model_id=stabilityai/stable-diffusion-xl-base-1.0 --revision=main --load
```

### sd2 with unet graph
run torch eager
```
python examples/text_to_image_sdxl_fp16_with_oneflow_compile.py
```

run oneflow graph
```
python examples/text_to_image_sdxl_fp16_with_oneflow_compile.py --compile
```

run oneflow graph save
```
python examples/text_to_image_sdxl_fp16_with_oneflow_compile.py --compile --save
```

run oneflow graph load
```
python examples/text_to_image_sdxl_fp16_with_oneflow_compile.py --compile --load
```
@strint strint closed this Aug 29, 2023
auto-merge was automatically disabled August 29, 2023 12:38

Pull request was closed

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Reviewers

@hjchen2 hjchen2 hjchen2 approved these changes

@liujuncheng liujuncheng Awaiting requested review from liujuncheng

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

Assignees
No one assigned
Labels
automerge enhancement feature graph graph mode
Projects
None yet
Milestone
No milestone
Development

Successfully merging this pull request may close these issues.
4 participants
@strint @hjchen2 @oneflow-ci-bot @jackalcooper