Skip to content

Releases: garry1ng/pytorch0.4.1_cuda11

.whl file

19 Jun 17:51
Compare
Choose a tag to compare

Tested on Ubuntu 18.04 (GLIBC 2.27), CUDA 11.1 cuda_11.1.0_455.23.05_linux.run with cudnn 8.0.5 cudnn-11.1-linux-x64-v8.0.5.39.tgz. You should set up your CUDA enviroment and conda prerequisites (numpy, mkl, ..etc) by yourself. RTX3090 should work well on this wheel file.
Note that your GLIBC version shoud be above 2.27.

Tips:

torchvision==0.2.2.post3