search

chengzeyi / stable-fast

Best inference performance optimization framework for HuggingFace Diffusers on NVIDIA GPUs.
MIT License
1.06k stars 60 forks source link

Please build release wheels for arch 8.6, 8.9 and arch 9.0 #89

Closed jon-chuang closed 6 months ago

jon-chuang commented 6 months ago

It may slow down wheel build but it should be better perf for newer architectures.

current: TORCH_CUDA_ARCH_LIST: "6.0 6.1 7.0 7.5 8.0+PTX"