Generate: remove near-duplicate sample/greedy copy #30773
Merged
Conversation
This file contains 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
|
The docs for this PR live here. All of your documentation changes will be reflected on that endpoint. The docs are available until 30 days after the last update. |
gante
commented
May 13, 2024
Comment on lines
+2254
to
+2266
| logger.warning_once( | ||
| "Calling `._greedy_search()` directly is deprecated and will be removed in v4.42. Use `._sample()` " | ||
| "instead, passing the same arguments." | ||
| ) | ||
| return self._sample( | ||
| input_ids=input_ids, | ||
| logits_processor=logits_processor, | ||
| stopping_criteria=stopping_criteria, | ||
| generation_config=generation_config, | ||
| synced_gpus=synced_gpus, | ||
| streamer=streamer, | ||
| **model_kwargs, | ||
| ) |
There was a problem hiding this comment.
There is no hard need for deprecations (these are private fns), but I know they are used in other internal libraries -- this should make the transition smoother!
| next_token_scores, max(2, 1 + n_eos_tokens) * num_beams, dim=1, largest=True, sorted=True | ||
| ) | ||
| n_tokens_to_keep = max(2, 1 + n_eos_tokens) * num_beams | ||
| if do_sample: |
There was a problem hiding this comment.
very nice thaat all the do_sample flag expose the real changed / diff between the two!
ArthurZucker
approved these changes
May 13, 2024
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Outdated suggestions cannot be applied.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?
We have two pairs of functions that are near copy of each other. These pairs are exactly the same, except for the token selection part (argmax vs multinomial sampling):
_greedy_search/_sample_beam_search/_beam_sampleThis PR keeps the most popular version of each pair (
_sampleand_beam_search), making the appropriate changes to preserve all functionality.