Fix Llama4 offset #37414
Merged
Fix Llama4 offset #37414
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
|
Hi 👋, thank you for opening this pull request! The pull request is converted to draft by default. The CI will be paused while the PR is in draft mode. When it is ready for review, please click the |
|
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. |
ArthurZucker
approved these changes
Apr 10, 2025
ArthurZucker
pushed a commit
that referenced
this pull request
Apr 10, 2025
* add +1 * Update modeling_llama4.py
cyr0930
pushed a commit
to cyr0930/transformers
that referenced
this pull request
Apr 18, 2025
* add +1 * Update modeling_llama4.py
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Outdated suggestions cannot be applied.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?
See title. It's the same as before + 1. Indeed, we need to add the +1 to match correctly the positions, it's also what is done in gemma2 here https://github.com/huggingface/transformers/blob/main/src/transformers/models/gemma2/modeling_gemma2.py#L912-L913 as it uses the length (which is equal to cache_position[-1] + 1).
I used
first_cache_position - attention_chunk_size + 1here as I feel it simpler to understand when looking at the cache code (see the 1-offset here https://github.com/huggingface/transformers/blob/main/src/transformers/cache_utils.py#L1921-L1922), but it is striclty equal. I.e. we always havemax(first_cache_position - attention_chunk_size + 1, 0) == max(last_cache_position + 1 - key_length, 0)