Fix: Make sft script work when chat template is None #3995
Closed
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
kashif
reviewed
Sep 2, 2025
| @@ -123,7 +126,7 @@ def main(script_args, training_args, model_args, dataset_args): | |||
| # Set default chat template if needed | |||
| if tokenizer.chat_template is None: | |||
| # TODO: source should be passed as an argument | |||
| model, tokenizer = clone_chat_template(model, tokenizer, "Qwen/Qwen3-0.6B") | |||
| model, tokenizer, _added_tokens = clone_chat_template(model, tokenizer, "Qwen/Qwen3-0.6B") | |||
There was a problem hiding this comment.
Suggested change
| model, tokenizer, _added_tokens = clone_chat_template(model, tokenizer, "Qwen/Qwen3-0.6B") | |
| model, tokenizer, _ = clone_chat_template(model, tokenizer, "Qwen/Qwen3-0.6B") |
There was a problem hiding this comment.
I wanted to do this as well, but then I made it more explicit to the users.
There was a problem hiding this comment.
as you like... ruff etc. might complain... i think this and the Optional should be the only change in this PR
qgallouedec
reviewed
Sep 2, 2025
|
|
||
| from accelerate import logging | ||
| from datasets import load_dataset | ||
| from transformers import AutoConfig, AutoModelForCausalLM, AutoTokenizer |
There was a problem hiding this comment.
I think you can revert this change
qgallouedec
reviewed
Sep 2, 2025
Comment on lines
-164
to
-168
| if subparsers is not None: | ||
| parser = subparsers.add_parser("sft", help="Run the SFT training script", dataclass_types=dataclass_types) | ||
| else: | ||
| parser = TrlParser(dataclass_types) | ||
| return parser |
There was a problem hiding this comment.
IMO, using a consistent indent level for the return is a better practice
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.
What does this PR do?
This pull request makes improvements to the
trl/scripts/sft.pyscript, primarily focusing on import optimization, correcting lazy importing, and minor API adjustments for clarity and correctness. The most significant changes are grouped below:Import optimization and lazy loading:
AutoModelForCausalLMat the top of the file and moved it inside the relevant code path in themainfunction to load it lazily.API and function signature improvements:
_added_tokensas the third argument becauseclone_chat_templatereturns three values, not two.make_parserfunction to make the type hint for thesubparsersaccurate and simplified the logic for returning the correct parser instance.Before submitting
Pull Request section?
to it if that's the case.
Who can review?
Anyone in the community is free to review the PR once the tests have passed. Feel free to tag
members/contributors who may be interested in your PR.