fix(distill): shared student_t5_config for sequence + logit paths - #174
Merged
Conversation
The logit path's raw T5Config(**student_config) ignored the spec's enc_layers/dec_layers vocabulary (T5Config defaults won: a 6/6 student from a 6/4 spec) — the Hebrew layerdrop run crashed in layer_drop_state with IndexError and would otherwise have silently trained wrong depths. student_t5_config() is now shared by the sequence and logit paths.
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Why
The Hebrew layerdrop launch crashed: layer_drop_state IndexError at
the decoder mapping. Root cause: the logit path built the student with
T5Config(**student_config), which does not understand enc_layers /
dec_layers — T5Config defaults silently produced a 6/6-depth student
from the 6/4 spec, and the verbatim layer copy (4 pretrained decoder
layers) indexed out of range. A 6/6 student would also have been the
wrong experiment entirely.
What
translation (byte-model constants included)
translation and defaults