Repository navigation
feat(pretrain): add torch_compile_args for SSL pretraining - #1002
IgorSusmelj wants to merge 3 commits into
Conversation
Codex Review SummaryThis comment shows the latest Codex review activity on this pull request.
ℹ️ About Codex in GitHubYour team has set up Codex to review pull requests in this repo. Reviews are triggered when you
Codex reacts with 👀 while any review is running, comments if it has suggestions, and reacts with 👍 once all reviews finish with no findings. |
![chatgpt-codex-connector[bot]](https://avatars.githubusercontent.com/in/1144995?s=60&v=4){kind=small}
There was a problem hiding this comment.
💡 Codex Review
Here are some automated review suggestions for this pull request.
Reviewed commit: 7c6384501b
ℹ️ About Codex in GitHub
Codex has been enabled to automatically review pull requests in this repo. Reviews are triggered when you
- Open a pull request for review
- Mark a draft as ready
- Comment "@codex review".
If Codex has suggestions, it will comment; otherwise it will react with 👍.
When you sign up for Codex through ChatGPT, Codex can also answer questions or update the PR, like "@codex address that feedback".
7c63845 to
ebc02f3
Compare
|
/review |
train_object_detectionetc.), but pretraining spends essentially all its time in backbone feature extraction too. This adds the same option topretrain()so long SSL runs can benefit as well.torch_compile_argsargument onpretrain()and the CLI, backed by the existingTorchCompileArgs. It compiles the backbone feature extraction in the train/val steps and falls back to eager execution if compilation fails. Off by default, so nothing changes for existing users. Cross-resolution input changes trigger recompilation, so it is best used with a fixed crop size (as in pretraining).tests/_commands/suite green (390 passed), ruff and mypy clean.