Repository navigation
feat: 한국어 파인튜닝 엔진을 8bit 변환본으로 교체 (4.7GB → 2.5GB) - #60
Merged
Merged
Conversation
- TeamUNIVA/qwen3_asr_1.7b_ko를 mlx_audio.convert로 8bit 변환해 voyagar/Qwen3-ASR-1.7B-ko-mlx-8bit(공개, Apache-2.0)로 올리고 등록 목록(asr-ko-univa)을 그 저장소로 교체 - 같은 5분 구간 비교: 원본과 출력 차이 CER 2.5, 처리 시간 약 30% 단축 - 문서·면제 기록·ARCHITECTURE v3.78·미리보기 데이터·테스트 갱신 Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
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
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
배경
0.5.2에 올린
asr-ko-univa는 허브 원본 bf16(4.7GB)을 그대로 등록한 것이었다. mlx 8bit로 변환해 소유자 계정에 올리고 등록 목록을 교체한다.변경
voyagar/Qwen3-ASR-1.7B-ko-mlx-8bit(공개, Apache-2.0, 커밋8fcc32c4)에 변환본 업로드:mlx_audio.convert -q --q-bits 8 --q-group-size 64 --q-mode affine(표준 엔진과 같은 설정, 가중치 크기도 같은 2,463,307,541)models_registry.py의asr-ko-univa를 그 저장소·리비전·파일 크기로 교체(ID는 그대로라 사용자가 고른 엔진 설정은 유지)간이 점검 (같은 회의 5분 구간, 11청크)
테스트
tractus models download asr-ko-univa→ 허브에서 받아installed(크기 일치)qwen3-asr-1.7b-ko-bf16폴더는 등록 밖이 되어models prune으로 지운다🤖 Generated with Claude Code