Stream-Omni enables seamless interactions across text, vision, and speech using a large language model. This repository includes the model, datasets, and tools for developers to explore multimodal ...