Apparently not. See https://github.com/lifeiteng/OmniSenseVoice/blob/main/src/om.... See also FunASR running SenseVoice but using Kaldi for speaker identification https://github.com/modelscope/FunASR/blob/cd684580991661b9a0...