ascendai/mindspore repository overview

⁠MindSpore

⁠Quick Reference

• MindSpore is maintained by the MindSpore Community⁠

• Where to get help

  • MindSpore Documentation⁠
  • MindSpore Community⁠
  • Issue Feedback⁠

⁠MindSpore

MindSpore is a new open source deep learning training/inference framework that could be used for mobile, edge and cloud scenarios. MindSpore is designed to provide development experience with friendly design and efficient execution for the data scientists and algorithmic engineers, native support for Atlas AI processor, and software hardware co-optimization. At the meantime MindSpore as a global AI open source community, aims to further advance the development and enrichment of the AI software/hardware application ecosystem.

⁠Supported Tags and Dockerfile Usage

⁠Tag Specification

Tags follow this format:

<MindSpore Version>-<CANN Version>-<Hardware Info (Chip)>-<Operating System>-<Python Version>

⁠Tag

• 2.10.0-cann9.1.0-910b-ubuntu22.04-py3.12
• 2.10.0-cann9.1.0-910b-openeuler24.03-py3.12
• 2.10.0-cann9.1.0-a3-ubuntu22.04-py3.12
• 2.10.0-cann9.1.0-a3-openeuler24.03-py3.12

See dockerfile:dockerfile⁠

⁠Building Arguments

⁠Image Repository Address

MindSpore Atlas images are hosted on Huawei Cloud SWR image repository:

swr.cn-south-1.myhuaweicloud.com/mindspore/mindspore

Full Image Example:

swr.cn-south-1.myhuaweicloud.com/mindspore/mindspore:2.10.0-cann9.1.0-910b-ubuntu22.04-py3.12

⁠Quick Start

⁠Building MindSpore Image

Taking the construction of the 2.10.0-cann9.1.0-910b-ubuntu22.04-py3.12 image as an example:

docker build \
    --build-arg CANN_VERSION=9.1.0 \
    --build-arg CHIP_ARCH=910b \
    --build-arg OS_SYSTEM=ubuntu22.04 \
    --build-arg PY_VERSION=py3.12 \
    --build-arg MINDSPORE_VERSION=2.10.0 \
    --build-arg PIP_INDEX_URL=https://mirrors.huaweicloud.com/repository/pypi/simple \
    -t mindspore:2.10.0-cann9.1.0-910b-ubuntu22.04-py3.12 \
    -f Dockerfile .

⁠Running MindSpore Container

docker run \
    --privileged \
    --name mindspore_container \
    --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 mindspore:{tag} bash

Where:

• {tag} is the tag specified during the MindSpore image build, for example, mindspore:2.10.0-cann9.1.0-910b-ubuntu22.04-py3.12.

⁠How to Extend for Custom Development

# Use MindSpore image as base image, add user software
FROM swr.cn-south-1.myhuaweicloud.com/mindspore/mindspore:2.10.0-cann9.1.0-910b-ubuntu22.04-py3.12

RUN apt update -y && \
    apt install -y gcc g++

# Install additional dependencies
RUN pip install pandas scikit-learn

# Copy user code
WORKDIR /workspace
COPY . /workspace

CMD ["python", "train.py"]

⁠Supported Hardware

⁠License

View the license information⁠ for MindSpore included in these images.

As with all container images, pre-installed software packages (Python, system libraries, etc.) may be subject to their respective licenses.

⁠Image Usage Responsibility

The released Atlas software images are community versions, not intended for commercial use, and serve only as references for production practices.