Please help me,i have a problem about cuda and pytorch version
kunzaiyy opened this issue · comments
After i installed cuda( conda install pytorch torchvision cudatoolkit=9.0 -c pytorch),when i ran python generate.py configs/demo.yaml,i got this problem::
RuntimeError: cublas runtime error : the GPU program failed to execute at /opt/conda/conda-bld/pytorch_1556653183467/work/aten/src/THC/THCBlas.cu:259
I inquired from Google and they told me that cuda9.0 is not supported by the graphics card, but when I applied cuda10.0, this would happen：：
ImportError: /home/kun/anaconda3/envs/mesh_funcspace/lib/python3.6/site-packages/torch/lib/libmkldnn.so.0: undefined symbol: cblas_sgemm_alloc
So,how did you run this code successfully，please tell me，my system version is ubuntu18.04,thank you very much!!
I also have the same problem.
solved by run command "conda install mkl -c anaconda"