2 ms·

Is your work open source?

• AndreSlavescu 10mo ago
  At the moment, no unfortunately. However, to my recent knowledge of open source alternatives, the vLLM team published a separate repository for omni models now: https://github.com/vllm-project/vllm-omni https://github.com/vllm-project/vllm-omni I have not yet tested out if this does full speech to speech, but this seems like a promising workspace for omni-modal models.