TOOLORA / Local reading edition
Fun-ASR
End-to-end ASR models: Nano focuses on Chinese, English, Japanese and Chinese dialects; MLT-Nano provides multilingual transcription through FunASR workflows.
Overview
Fun-ASR is Tongyi Lab’s recognition-model family, distinct from the multi-model FunASR toolkit. Nano and MLT-Nano have different language scopes, with MLT-Nano targeting thirty-one languages. Direct, Transformers and FunASR workflows provide transcription, with hotwords and streaming depending on the runtime. Timestamps depend on the downloaded checkpoint, not its name alone; diarization needs additional models. Main code and the reviewed official Nano card declare Apache 2.0.
Official sources and references
- Official websitehttps://github.com/QwenAudio/Fun-ASR
- Documentationhttps://huggingface.co/FunAudioLLM/Fun-ASR-Nano-2512
Descriptions summarize public information; consult official websites for features, pricing and regional availability.