gpt-omni/mini-omni

VoiceAssistant-400K

Closed this issue · 3 comments

UestcJay commented

Thanks for your great work! Any plan to open the data production process of VoiceAssistant-400K?

mini-omni commented

Hi, for the data processing, we use cosyvoice for TTS (input audio) and SNAC for output audio encoding, you may refer to these projects for more info.

UestcJay commented

thanks for your reply! I don't mean the production of the audio, but the production of these single or multi-turn conversations?

mini-omni commented

we use open-source datasets, you may refer to tech report the dataset sources.