Searching protocol for "omni-models"
Multi-modal AI for text, image, audio, and video.
Reduce VRAM usage and speed up vLLM-Omni.