Skip to content

Latest commit

 

History

History
executable file
·
19 lines (13 loc) · 838 Bytes

README.md

File metadata and controls

executable file
·
19 lines (13 loc) · 838 Bytes

docker-cudaml

Repository which has some base images for running CUDA and cuDNN on Intel and ARM architectures.

CUDA Images

If you want to use an NVIDIA GPU, You can use the following two images as the basis for your own images:

  • ghcr.io/mutablelogic/cuda-dev:1.0.2 - This image is based on Ubuntu 22.04 and includes the 12.6 CUDA toolkit and compiler build tools
  • ghcr.io/mutablelogic/cuda-rt:1.0.2 - This image is based on Ubuntu 22.04 and includes the 12.6 CUDA runtime libraries.

When running a runtime container, then install the NVIDIA Container Toolkit first. Then you can run the container with the following command:

docker run \
  --name <name> --rm \
  --runtime nvidia --gpus all 
  <image> <arguments>