Set Gemma 3 4B max_position_embeddings to 128K#95
Open
paper-turtle wants to merge 1 commit intogoogle:mainfrom
Open
Set Gemma 3 4B max_position_embeddings to 128K#95paper-turtle wants to merge 1 commit intogoogle:mainfrom
paper-turtle wants to merge 1 commit intogoogle:mainfrom
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
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
max_position_embeddingsis missing from the Gemma 3 4B config. It defaults to8192and the model will error if you pass more than2*8192tokens.This PR sets
max_position_embeddingsto131072(128*1024) to match the 128K context length stated in the paper.Note: In the Kaggle checkpoints for
gemma-3-4b-itandgemma-3-4b-pt, thelocal_freqs_cisandglobal_freqs_cistensors currently have shape(16384, 128). The checkpoints will need to be updated, otherwiseload_state_dictwill fail due to the shape mismatch.