I used export TORCH_CUDA_ARCH_LIST="6.0" to reduce the computing power of RTX3070, and successfully installed a series of environments such as cuda10.2+pytorch1.5, but it took 10 minutes to create a model when running demos/face_demo.py. The content can't come out, why is this?
I used export TORCH_CUDA_ARCH_LIST="6.0" to reduce the computing power of RTX3070, and successfully installed a series of environments such as cuda10.2+pytorch1.5, but it took 10 minutes to create a model when running demos/face_demo.py. The content can't come out, why is this?