fix(distill): logit-KD student vocab matches teacher - #175
Merged
Conversation
The KL between teacher and student distributions requires identical vocab; student_t5_config hardcoded the sequence path's 259-token byte-table, so the Hebrew layerdrop crashed in F.kl_div (384 vs 259). vocab_size is now spec-configurable (259 default preserved); the Hebrew layerdrop spec pins 384 to match the ByT5 s46 teacher.
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
F.kl_div crashed: teacher logits vocab 384 vs student 259. vocab_size now spec-configurable with the sequence-path default preserved; Hebrew layerdrop spec pins 384. Suite 132 green.