@zyang-dev · User
LLM inference in C/C++
Optimized primitives for collective multi-GPU communication
