1 paper
Yaochen Han, Ke Fan, Hongxu Jiang +5
High-performance GPU kernels are critical for reducing the exponentially growing computational costs of large language models (LLMs), but their development heavily relies on manual…