I have been trying to test the installation of the GPU-enabled Pytorch on Windows using Anaconda, and had the following observations:
Whether the version is Stable (1.9.1) or LTS(1.8.2),(conda install pytorch torchvision torchaudio cudatoolkit=10.2 -c pytorch )I have to choose CUDA 10.2 and the test of CUDA will pass
Import torch
Torch.cuda.is_available()
True
if I choose CUDA11.1, the installation will fail to pass the following test
import torch
torch.cuda.is_available()
False
Here is the system information for the reference; I am not clear why I cannot use CUDA11.1 in the Pytorch installation, and what are the criteria to choose it?
Here is the command I used to install Pytorch with CUDA11.1
conda install pytorch torchvision torchaudio cudatoolkit=11.1 -c pytorch -c conda-forge
Here is the information I run the env collection script:
Collecting environment information...
PyTorch version: 1.9.0+cpu
Is debug build: False
CUDA used to build PyTorch: Could not collect
ROCM used to build PyTorch: N/A
OS: Microsoft Windows 10 Enterprise
GCC version: (MinGW.org GCC-6.3.0-1) 6.3.0
Clang version: Could not collect
CMake version: version 3.20.2
Libc version: N/A
Python version: 3.8.12 | packaged by conda-forge | (default, Sep 29 2021, 19:15:57) [MSC v.1916 64 bit (AMD64)] (64-bit runtime)
Python platform: Windows-10-10.0.19043-SP0
Is CUDA available: False
CUDA runtime version: 11.2.67
GPU models and configuration: Could not collect
Nvidia driver version: Could not collect
cuDNN version: Could not collect
HIP runtime version: N/A
MIOpen runtime version: N/A
Versions of relevant libraries:
[pip3] botorch==0.5.0
[pip3] gpytorch==1.5.0
[pip3] numpy==1.20.3
[pip3] torch==1.9.0
[pip3] torchaudio==0.9.1
[pip3] torchvision==0.10.1
[conda] blas 2.111 mkl conda-forge
[conda] blas-devel 3.9.0 11_win64_mkl conda-forge
[conda] cudatoolkit 11.1.1 heb2d755_9 conda-forge
[conda] libblas 3.9.0 11_win64_mkl conda-forge
[conda] libcblas 3.9.0 11_win64_mkl conda-forge
[conda] liblapack 3.9.0 11_win64_mkl conda-forge
[conda] liblapacke 3.9.0 11_win64_mkl conda-forge
[conda] mkl 2021.3.0 hb70f87d_564 conda-forge
[conda] mkl-devel 2021.3.0 h57928b3_565 conda-forge
[conda] mkl-include 2021.3.0 hb70f87d_564 conda-forge
[conda] numpy 1.21.2 py38h089cfbf_0 conda-forge
[conda] pytorch 1.9.1 py3.8_cuda11.1_cudnn8_0 pytorch
[conda] torchaudio 0.9.1 py38 pytorch
[conda] torchvision 0.10.1 py38_cu111 pytorch