tkrmagid/javis_bot
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases 20 Wiki Activity

perf(stt): default Whisper model small -> medium for better Korean accuracy

Browse Source 
Whisper 'small' frequently mis-transcribes Korean (esp. the wake word 자비스:
'아비스', '자 비싸'). STT is only ~0.1s warm so there is ample latency headroom;
'medium' is far more accurate and fits VRAM alongside qwen2.5:3b on the 8GB GPU.

Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
This commit is contained in:
javis-bot
2026-06-13 22:03:18 +09:00
parent 3a4776c709
commit 4e446c1d7c
1 changed files with 1 additions and 1 deletions
Download Patch File Download Diff File Expand all files Collapse all files

2
docker-compose.yml
View File

@@ -64,7 +64,7 @@ services:
OLLAMA_BASE_URL: http://ollama:11434
OLLAMA_CHAT_MODEL: ${OLLAMA_CHAT_MODEL:-qwen2.5:3b}
OLLAMA_EMBED_MODEL: ${OLLAMA_EMBED_MODEL:-nomic-embed-text}
WHISPER_MODEL: ${WHISPER_MODEL:-small}
WHISPER_MODEL: ${WHISPER_MODEL:-medium}
WHISPER_DEVICE: ${WHISPER_DEVICE:-cuda}
WHISPER_COMPUTE_TYPE: ${WHISPER_COMPUTE_TYPE:-float16}
# Optional single-language lock for replies (empty = user's own language).