Chorus v1: Overlapping Speech Transcription
Posted by TrelisResearch@reddit | LocalLLaMA | View on Reddit | 3 comments
New open weights model allowing for multi-speaker transcription using a single model. PyTorch and here ggml weights and a whisper-cli patch provided.
uutnt@reddit
The link is down.
TrelisResearch@reddit (OP)
thanks, sorry made it public now. See also the main repo on hf here: https://huggingface.co/Trelis/Chorus-v1
ComplexType568@reddit
It isn't anymore, as far as I can tell