Back to search
Top Contributors
77
commits
Multi-source transcript merging inspired by textual criticism — LLM adjudicates multiple Whisper, YouTube captions & external transcripts for higher quality. Includes speaker diarization and summarization.
Stars
18
Forks
1
Watchers
18
Open Issues
1
Overall repository health assessment
No package.json found
This might not be a Node.js project
77
commits
Pipeline improvements: skip redundant work, fix DRY violations, unify ASR terminology
7b8f479View on GitHubRemove backward compat: legacy whisper_* files trigger migration error
7693cd2View on GitHubChange default to parakeet only (24.7% WER, best single model)
612245eView on GitHubFilter ensemble to requested models only, ignore extras on disk
b86c132View on GitHubAdd tests for coverage gaps and cross-stage pipeline transitions
00fbfebView on GitHubDRY refactor: extract 6 shared helpers to eliminate duplicated patterns
cb633d8View on GitHubUnify ASR naming: replace Whisper-centric naming with generic asr_* convention
9105a6fView on GitHubAdd Collins cross-arch ensemble results to experiment docs
2f02b8eView on GitHubFix ensemble base selection bug, update quality ranks, add tests
30f6099View on GitHub