AutoMegaKernel: Compiling a LLM into a single CUDA kernel
Posted 2 hours ago by
OsamaJaber
3
points
https://arxiv.org/abs/2606.09682
0
comments