search

FunAudioLLM / SenseVoice

Multilingual Voice Understanding Model
https://funaudiollm.github.io/
Other
3.34k stars 307 forks source link

sensevoice-small模型推理速度与报告不符 #42

Closed yzslry closed 3 months ago

yzslry commented 3 months ago

使用长度为9.42s的音频实际测试，encoder的推理时间就需要0.5s左右，开源的模型是报告中10s音频推理仅需70ms的模型吗？

LauraGPT commented 3 months ago

Please upload you wav and logfile.

MonolithFoundation commented 3 months ago

别用，效果奇差。和whisper差远了。