VLLM-Omni: A framework for efficient model inference with Omni-modality models github.com 1 points by zyh888 4 hours ago
everlier 3 hours ago I feel like there's a story behind multimodal architectures not being integrated into the main package by default
I feel like there's a story behind multimodal architectures not being integrated into the main package by default