Researchers from NVIDIA, CMU and the University of Washington released 'FlashInfer': a kernel library that provides next-generation kernel implementations for LLM inference and serving 01/05/2025
