Skip to content

mindspore-ai/xai

Repository files navigation

Explainable AI (XAI)

查看中文

What is Explainable AI (XAI)

XAI is an explainable AI toolbox based on MindSpore. Currently, most deep learning models are black-box models with good performance but poor explainability. XAI aims to provide users decision explanations of the model, helps users better understand the model, trust the model, and improve the model when an error occurs. Besides a variety of explanation methods, we also provide a set of evaluation methods to evaluate the explanation methods from various dimensions. It helps users compare and select the explanation methods that are most suitable for a particular scenario.

System Architecture

sys_arch

Installation

System Requirements

  • OS: EulerOS-aarch64, CentOS-aarch64, CentOS-x86, Ubuntu-aarch64 or Ubuntu-x86
  • Device: Ascend 910 or GPU CUDA 10.1, 11.1
  • Python 3.7.5 or 3.9.0
  • MindSpore 1.7 or above

Installing by pip

The simplest way to do it:

pip install mindspore-xai

Or you may install with a downloaded .whl package from MindSpore XAI download page:

pip install mindspore_xai-{version}-py3-none-any.whl

Installing from Source Code

  1. Download source code from gitee.com:
git clone https://gitee.com/mindspore/xai.git
  1. Install the dependency python modules:
cd xai
pip install -r requirements.txt
  1. Install the XAI module from source code:
python setup.py install
  1. Optionally, you may build a .whl package for installation without step 3:
bash package.sh
pip install output/mindspore_xai-{version}-py3-none-any.whl

Installation Verification

Upon successful installation, importing 'mindspore_xai' module in python will cause no error:

import mindspore_xai
print(mindspore_xai.__version__)

Quick Start and API References

For the quick start guide and API references, please checkout XAI Homepage.

Community

Governance

Checkout how MindSpore Open Governance works

Contributing

Welcome contributions. See our Contributor Wiki for more details.

License

Apache License 2.0