search

egoist / whispo

AI-powered dictation tool
GNU Affero General Public License v3.0
322 stars 16 forks source link

The option to set the model prompt and model name is required #2

Open BryceWG opened 3 weeks ago

BryceWG commented 3 weeks ago

Allowing configuring the model prompt and the model request name to better support more models is very important, I wish my demand will be satisfied.

egoist commented 3 weeks ago

What do you need other than the whisper model?

BryceWG commented 3 weeks ago

I like the FunAudioLLM/SenseVoiceSmall model on the siliconflow platform, which is very fast, accurate and suitable for Chinese. And the platform will add more ASR model this month. The post-processing also needs an option to choose a text model, and I don't know what's the default post-processing model