[refactor] support compile model and loss separately #1608
Merged
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
meta-cla
bot
added
the
CLA Signed
wwwjn
reviewed
Aug 20, 2025
| if parallel_dims.tp_enabled: | ||
| if ( | ||
| job_config.parallelism.enable_async_tensor_parallel | ||
| and not job_config.training.compile | ||
| and not model_compile_enabled | ||
| ): | ||
| raise RuntimeError("Async TP requires --training.compile") |
There was a problem hiding this comment.
Need to fix the error message
| if parallel_dims.tp_enabled: | ||
| if ( | ||
| job_config.parallelism.enable_async_tensor_parallel | ||
| and not job_config.training.compile | ||
| and not model_compile_enabled | ||
| ): | ||
| raise RuntimeError("Async TP requires --training.compile") |
There was a problem hiding this comment.
Need to fix the error message as well
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Creating a new field in
JobConfig, with the default beingThis way we get to compile loss separately to get memory reduction, even when the model is not ready to be compiled.
This PR also applies loss compilation to DeepSeek 16B and 671B.