I have been wanting to move my pytorch project to the GPUs for training. I am working inside a docker container. Thus, I tested torch.cuda.is_available() wich returned False .

This is what my dockerfile looks like.

FROM mcr.microsoft.com/devcontainers/python:0-3.11

ENV PATH=/usr/local/cuda/bin:${PATH}

ENV LD_LIBRARY_PATH=/usr/local/cuda/lib64:${LD_LIBRARY_PATH}

WORKDIR /app

CMD ["/bin/bash"]

I tried with nvidia-smi which gave me this

+-----------------------------------------------------------------------------+
| NVIDIA-SMI 470.161.03   Driver Version: 470.161.03   CUDA Version: 11.4     |
|-------------------------------+----------------------+----------------------+
| 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 A100-PCI...  Off  | 00000000:01:00.0 Off |                    0 |
| N/A   29C    P0    31W / 250W |      4MiB / 40536MiB |      0%      Default |
|                               |                      |             Disabled |
+-------------------------------+----------------------+----------------------+
|   1  NVIDIA A100-PCI...  Off  | 00000000:41:00.0 Off |                    0 |
| N/A   30C    P0    34W / 250W |      4MiB / 40536MiB |      0%      Default |
|                               |                      |             Disabled |
+-------------------------------+----------------------+----------------------+
|   2  NVIDIA A100-PCI...  Off  | 00000000:81:00.0 Off |                    0 |
| N/A   29C    P0    32W / 250W |      4MiB / 40536MiB |      0%      Default |
|                               |                      |             Disabled |
+-------------------------------+----------------------+----------------------+
|   3  NVIDIA A100-PCI...  Off  | 00000000:C1:00.0 Off |                    0 |
| N/A   29C    P0    33W / 250W |      4MiB / 40536MiB |      0%      Default |
|                               |                      |             Disabled |
+-------------------------------+----------------------+----------------------+
                                                                               
+-----------------------------------------------------------------------------+
| Processes:                                                                  |
|  GPU   GI   CI        PID   Type   Process name                  GPU Memory |
|        ID   ID                                                   Usage      |
|=============================================================================|
+-----------------------------------------------------------------------------+

But when trying nvcc --version it returns bash: nvcc: command not found . This issue does not arise when I type nvcc --version outside of my container. I thus assume, that it has something to do with my docker container.

Any tips what I might be doing wrong?