Model release · Xiaomi
Xiaomi-CocktailASR-1
- Model released
目标说话人 ASR SpeechLLM（鸡尾酒会场景）；Data2Vec2+PerceptionAdapter+Qwen3；Apache 2.0 开源；2026-09-11 小米官宣
Availability
Open weights (Apache 2.0): GitHub xiaomi-research/xiaomi-cocktailasr-1; Hugging Face Ease3/Xiaomi-CocktailASR-1
Source trace
Reviewed by Ailyre. Source evidence is presented separately from our summary.
Ease3/Xiaomi-CocktailASR-1 · Hugging Face
Hugging Face
Source checked:
Original ↗View archived evidence ↗
This legacy archive has no reviewed checksum; its body is not published.
