1 paper · 1 filter
Zhongzhen Wen, Shudi Shao, Zhong Li +4
The performance of deep learning models critically depends on efficient kernel implementations, yet developing high-performance kernels for specialized accelerators remains time-co…