fix: distill lint on main + Thai teacher recovery + parts-aware publish - #12
Merged
Conversation
added 4 commits
August 18, 2026 17:04
- ruff: unused val_path/epoch/math in modal_distill/modal_teacher_thai - modal_teacher_thai.py: recover the Thai umt5 teacher under transformers 5.14.1 — every secryst-saved umt5 artifact is unusable (5.15 save dropped the untied lm_head; verified by probes 2026-08-18). Same recipe as train_thai_combined.py: B-K base + 60K Kaikki+epitran. - decode_joined everywhere: 5.x batch_decode inserts spurious spaces between sentencepiece pieces — this mangled the first student's teacher labels; tha-g2p-small moves to run-002 (retrain pending) - publish_model.py: worktree-based branch handling (dirty trees no longer block publication), skip re-upload when assets already match
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Summary
tha-g2p-smallentry inmodal_export.pyMODELS (checkpointsecryst_thai_g2p_distill_small/run-001/beston secryst-checkpoints)evaluate_perinmodal_distill.py: teacher/student PER gate on the SAME harness as the source results — held-outthai-ipa/test.jsonl(1,219 Kaikki sentences), beam-4, corpus-level PER (train_thai_combined.py protocol); gate = teacher + 5ppmodal_distill.py(unusedval_path/epochflagged by ruff on main)Test plan
evaluate_per --spec tha-g2p-smallpasses before export runs