Fork自 PaddlePaddle / Paddle
* set auto to reduce core_avx/noavx.pyd size * set CUDA_ARCH_NAME in each case
拖放文件到此处或点击上传