Improve postprocessing efficiency #157
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
This comment was marked as outdated.
This comment was marked as outdated.
Sorry, something went wrong.
There was a problem hiding this comment.
Pull Request Overview
This PR refactors token postprocessing to improve performance by caching special tokens and vectorizing the assignment of character probabilities. Key changes include:
- Caching tokenizer special tokens to avoid repeated getattr calls.
- Replacing an iterative loop with vectorized indexing in token_to_char_probs.
- Updating function calls across multiple modules to use the new special_tokens parameter.
Reviewed Changes
Copilot reviewed 5 out of 5 changed files in this pull request and generated 1 comment.
Show a summary per file
| File | Description |
|---|---|
| wtpsplit/utils/init.py | Updated get_token_spans and token_to_char_probs for efficiency refactor. |
| wtpsplit/train/evaluate.py | Adjusted token_to_char_probs call to pass the cached special_tokens. |
| wtpsplit/evaluation/intrinsic_pairwise.py | Modified token_to_char_probs call to include the new special_tokens. |
| wtpsplit/evaluation/adapt.py | Updated token_to_char_probs calls with the new special_tokens parameter. |
| wtpsplit/init.py | Cached special_tokens in the initializer and updated usage accordingly. |
Comments suppressed due to low confidence (1)
wtpsplit/utils/init.py:445
- Consider updating the function docstring to specify that the special_tokens parameter should be a list of tokens, explaining its role in replacing repeated getattr calls.
def token_to_char_probs(text, tokens, token_logits, special_tokens, offsets_mapping):
Co-authored-by: Copilot <175728472+Copilot@users.noreply.github.com>
|
Thanks for the suggestions, looks all good to me! I will merge it. |
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.
This PR makes a couple of small refactors to improve performance with processing logits. On an L40S, this translates to a ~3x speed improvement.
get_token_spans,getattrcalls totokenizer.cls_token, tokenizer.sep_token, tokenizer.pad_tokendominate and are replaced with a one-time definition when the tokenizer is first instantiated.token_to_char_probs, assignment ofchar_probsis replaced with vectorized indexing instead of a for loop.Traces from pyinstrument:
Before:
After:
