Describe the bug
first, if you dont care the speed ,could use #142 ,modify the file "custom_os.py" device=torch.device(cpu),it could work for me.
and then i study why gpu get this error. i found that i couldnt build "bias_act_plugin",and i inspect the bias_act.cpp, it amazes me that gcc coundnt find includes(invovle"torch/extension.h,ATen/cuda/……")
i guess the error should because this .and originally, i think may that because i install torch by download wheel?
any idea please! i'm much appreciate.
Desktop (please complete the following information):
OS: Windows 11
PyTorch version 1.12.0
CUDA toolkit version 11.6
MVS 2022
NVIDIA driver version
GPU 1660ti
Additional context
Add any other context about the problem here.
Han-duoduo
commented
1 year ago
Describe the bug first, if you dont care the speed ,could use #142 ,modify the file "custom_os.py"
device=torch.device(cpu),it could work for me.and then i study why gpu get this error. i found that i couldnt build "bias_act_plugin",and i inspect the bias_act.cpp, it amazes me that gcc coundnt find includes(invovle"torch/extension.h,ATen/cuda/……")
i guess the error should because this .and originally, i think may that because i install torch by download wheel?
any idea please! i'm much appreciate.
Desktop (please complete the following information):
Additional context Add any other context about the problem here.