OpenMOSS / MOSS-VL

An open-weight 11B model series for long-form and real-time video understanding

xLLM-AI / xllm

A high-performance inference engine for LLM, VLM, DiT and REC models, optimized for diverse AI accelerators. It is hosted in OpenAtom Foundation.

brontoguana / krasis

Krasis is a Hybrid LLM runtime which focuses on efficient running of larger models on consumer grade VRAM limited hardware