问题:
RuntimeError: CUDA error: no kernel image is available for execution on the device
CUDA kernel errors might be asynchronously reported at some other API call,so the stacktrace below might be incorrect.
For debugging consider passing CUDA_LAUNCH_BLOCKING=1.
Solution idea,
I thought it was a video memory problem, but after debugging, I found that it didn’t seem to be
set according to the prompts
os.environ['CUDA_LAUNCH_BLOCKING'] = '1'
It is not right. Later, I checked on the Internet and felt that the CUDA computing power did not match when compiling.
Final solution:
Since I use 1080Ti, the corresponding computing power 61
is added in the setup25 line
"-gencode=arch=compute_61,code=sm_61",
Recompile and there is no problem.