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.

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

Use SGD for list of models from PyTorch. #6324

Merged
merged 4 commits into from
Jan 20, 2024

Conversation

ysiraichi
Copy link
Collaborator

This PR replaces Adam for SGD optimizer for a few models that are tight on memory. This list was taken from PyTorch's main repository:

https://github.com/pytorch/pytorch/blob/eba5d5485dd3232737ac789f1614f4cbc09220a4/benchmarks/dynamo/common.py#L214-L252

@@ -29,6 +29,27 @@
"detectron2_fcos_r_50_fpn",
}

# torchbench models that might OOM using Adam.
# This list was extracted from PyTorch's repository: benchmarks/dynamo/common.py
Copy link
Collaborator

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

So in the link it is mentioned that they use Adam optimiser in the CI. What optimiser is used in the official PyTorch HUD? I'm asking because we want to be as close as possible.

Copy link
Collaborator Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Good catch. So, maybe we want to get the previous set: BENCHMARK_USE_SGD, which is used outside CI.

@ysiraichi ysiraichi force-pushed the ysiraichi/fix-model-optimizer branch from cfae238 to b6d6acd Compare January 20, 2024 13:50
@ysiraichi ysiraichi merged commit a7918a7 into master Jan 20, 2024
18 checks passed
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

Successfully merging this pull request may close these issues.

3 participants