Quick reference

• The official PyTorch docker image.

• Maintained by: openEuler CloudNative SIG.

• Where to get help: openEuler CloudNative SIG, openEuler.

PyTorch | openEuler

Current PyTorch docker images are built on the openEuler. This repository is free to use and exempted from per-user rate limits.

PyTorch is a Python package that provides two high-level features:

• Tensor computation (like NumPy) with strong GPU acceleration
• Deep neural networks built on a tape-based autograd system

You can reuse your favorite Python packages such as NumPy, SciPy, and Cython to extend PyTorch when needed.

Learn more about on PyTorch Website.

Supported tags and respective Dockerfile links

The tag of each pytorch docker image is consist of the complete software stack version. The details are as follows

Tag	Currently	Architectures
pytorch2.1.0.a1-cann7.0.RC1.alpha002-oe2203sp2	pyTorch 2.1.0.a1 with cann 7.0.RC1.alpha002 on openEuler 22.03-LTS-SP2	arm64
2.2.0-cann8.0.RC1-oe2203sp4	PyTorch 2.2.0 with CANN 8.0.RC1 on openEuler 22.03-LTS-SP4	arm64,amd64

Usage

In this usage, users can select the corresponding {Tag} and container startup options based on their requirements.

• Pull the openeuler/pytorch image from docker

  docker pull openeuler/pytorch:{Tag}
  

• Start a pytorch instance

  docker run \
      --name my-pytorch \
      --device /dev/davinci1 \
      --device /dev/davinci_manager \
      --device /dev/devmm_svm \
      --device /dev/hisi_hdc \
      -v /usr/local/dcmi:/usr/local/dcmi \
      -v /usr/local/bin/npu-smi:/usr/local/bin/npu-smi \
      -v /usr/local/Ascend/driver/lib64/:/usr/local/Ascend/driver/lib64/ \
      -v /usr/local/Ascend/driver/version.info:/usr/local/Ascend/driver/version.info \
      -v /etc/ascend_install.info:/etc/ascend_install.info \
      -it openeuler/pytorch:{Tag} bash
  

• Container startup options

  Option	Description
  --name my-pytorch	Names the container my-pytorch.
  --device /dev/davinciX	NPU device, where X is the physical ID number of the chip, e.g., davinci1.
  --device /dev/davinci_manager	Davinci-related management device.
  --device /dev/devmm_svm	Memory management-related device.
  --device /dev/hisi_hdc	HDC-related management device.
  -v /usr/local/dcmi:/usr/local/dcmi	Mounts the host's DCMI .so and interface file directory /usr/local/dcmi to the container. Please modify according to actual situation.
  -v /usr/local/bin/npu-smi:/usr/local/bin/npu-smi	Mount the host npu-smi tool "/usr/local/bin/npu-smi" into the container. Please modify it according to the actual situation.
  -v /usr/local/Ascend/driver/lib64/:/usr/local/Ascend/driver/lib64/	Mounts the host directory /usr/local/Ascend/driver/lib64/driver to the container. Please modify according to the path where the driver's .so files are located.
  -v /usr/local/Ascend/driver/version.info:/usr/local/Ascend/driver/version.info	Mounts the host's version information file /usr/local/Ascend/driver/version.info to the container. Please modify according to actual situation.
  -v /etc/ascend_install.info:/etc/ascend_install.info	Mounts the host's installation information file /etc/ascend_install.info to the container.
  -it	Starts the container in interactive mode with a terminal (bash).
  openeuler/pytorch:{Tag}	Specifies the Docker image to run, replace {Tag} with the specific version or tag of the openeuler/pytorch image you want to use.

• View container running logs

  docker logs -f my-pytorch
  

• To get an interactive shell

  docker exec -it my-pytorch /bin/bash
  

Question and answering

If you have any questions or want to use some special features, please submit an issue or a pull request on openeuler-docker-images.