Fetching the paper…

Freeze-Omni: A Smart and Low Latency Speech-to-speech Dialogue Model with Frozen LLM · Around