Sm_86 is not compatible with current pytorch version

NVIDIA RTX A4000 with CUDA capability sm_86 is not compatible with the current PyTorch installation.
The current PyTorch install supports CUDA capabilities sm_37 sm_50 sm_60 sm_70.

Here is output of python -m torch.utils.collect_env

Collecting environment information...
PyTorch version: 1.11.0
Is debug build: False
CUDA used to build PyTorch: 11.3
ROCM used to build PyTorch: N/A

OS: Ubuntu 20.04.4 LTS (x86_64)
GCC version: (Ubuntu 9.4.0-1ubuntu1~20.04.1) 9.4.0
Clang version: Could not collect
CMake version: Could not collect
Libc version: glibc-2.31

Python version: 3.9.12 (main, Apr  5 2022, 06:56:58)  [GCC 7.5.0] (64-bit runtime)
Python platform: Linux-
Is CUDA available: False
CUDA runtime version: 11.3.109
GPU models and configuration: GPU 0: NVIDIA RTX A4000
Nvidia driver version: 512.59
cuDNN version: Probably one of the following:
HIP runtime version: N/A
MIOpen runtime version: N/A

Versions of relevant libraries:
[pip3] numpy==1.21.5
[pip3] torch==1.11.0
[pip3] torchaudio==0.11.0
[pip3] torchsummary==1.5.1
[pip3] torchvision==0.12.0
[conda] blas                      1.0                         mkl
[conda] cudatoolkit               11.3.1               h2bc3f7f_2
[conda] ffmpeg                    4.3                  hf484d3e_0    pytorch
[conda] mkl                       2021.4.0           h06a4308_640
[conda] mkl-service               2.4.0            py39h7f8727e_0
[conda] mkl_fft                   1.3.1            py39hd3c417c_0
[conda] mkl_random                1.2.2            py39h51133e4_0
[conda] numpy                     1.21.5           py39he7a7128_2
[conda] numpy-base                1.21.5           py39hf524024_2
[conda] pytorch                   1.11.0          py3.9_cuda11.3_cudnn8.2.0_0    pytorch
[conda] pytorch-mutex             1.0                        cuda    pytorch
[conda] torchaudio                0.11.0               py39_cu113    pytorch
[conda] torchsummary              1.5.1                    pypi_0    pypi
[conda] torchvision               0.12.0               py39_cu113    pytorch

Here is output of nvidia-smi

| NVIDIA-SMI 510.68.01    Driver Version: 512.59       CUDA Version: 11.6     |
| GPU  Name        Persistence-M| Bus-Id        Disp.A | Volatile Uncorr. ECC |
| Fan  Temp  Perf  Pwr:Usage/Cap|         Memory-Usage | GPU-Util  Compute M. |
|                               |                      |               MIG M. |
|   0  NVIDIA RTX A4000    On   | 00000000:65:00.0  On |                    0 |
| 41%   29C    P8    13W / 140W |    326MiB / 15352MiB |      2%      Default |
|                               |                      |                  N/A |

| Processes:                                                                  |
|  GPU   GI   CI        PID   Type   Process name                  GPU Memory |
|        ID   ID                                                   Usage      |
|  No running processes found                                                 |

I have tried uninstallation/installation so many times. cant figure out why is this happening?

1 Like

Based on the error message you’ve installed a PyTorch release with the CUDA10.2 runtime, which doesn’t match the posted environment. Uninstall all PyTorch installations and reinstall the PyTorch release with CUDA11 or create a new environment and reinstall it there.

1 Like

I have the same issue, did you solve the problem?

You might be running into the same issue of using the older CUDA 10.2 runtime.
Uninstall all other PyTorch binaries and install the PyTorch binary with CUDA11.3 or 11.6.

1 Like

Ok thank you Ill try

I tried to install torch==1.9.0 for CUDA = 11.6, but still getting the same error. Should I reinstall CUDA with version = 11.3?

torch==1.9.0 should not have supported CUDA 11.6 at all, so I’m unsure where you are getting the binary from. Use the latest stable release (1.12.1) with a supported CUDA runtime (e.g. 11.6).

Thanks, yea then I have to install the corresponding version of torchvision right?

Yes, I would keep all packages compatible to each other by installing the versions corresponding to the same releases.