kitstar / nni

An open source AutoML toolkit for neural architecture search and hyper-parameter tuning. https://microsoft.github.io/nni/

Neural Network Intelligence

NNI (Neural Network Intelligence) is a toolkit to help users run automated machine learning (AutoML) experiments. The tool dispatches and runs trial jobs generated by tuning algorithms to search the best neural architecture and/or hyper-parameters in different environments like local machine, remote servers and cloud.

Who should consider using NNI

• Those who want to try different AutoML algorithms in their training code (model) at their local machine.
• Those who want to run AutoML trial jobs in different environments to speed up search (e.g. remote servers and cloud).
• Researchers and data scientists who want to implement their own AutoML algorithms and compare it with other algorithms.
• ML Platform owners who want to support AutoML in their platform.

Install & Verify

Install through source code

• We only support Linux (Ubuntu 16.04 or higher) in our current stage.
• Run the following pip install in an environment that has python >= 3.5, git and wget.

    git clone -b v0.3 https://github.com/Microsoft/nni.git
    cd nni
    source install.sh

Verify install

• The following example is an experiment built on TensorFlow. Make sure you have TensorFlow installed before running it.
• Download the examples via clone the source code.

    cd ~
    git clone -b v0.3 https://github.com/Microsoft/nni.git

• Run the mnist example.

    nnictl create --config ~/nni/examples/trials/mnist/config.yml

• Wait for the message Info: Start experiment success! in the command line. This message indicates that your experiment has been successfully started. You can explore the experiment using the Web UI url.

    Info: Checking experiment...
    ...
    Info: Starting experiment...
    Info: Checking web ui...
    Info: Starting web ui...
    Info: Starting web ui success!
+   Info: Web UI url: http://127.0.0.1:8080   http://10.172.141.6:8080
+   Info: Start experiment success! The experiment id is LrNK4hae, and the restful server post is 51188.

Documentation

• Overview
• Get started

How to

• Installation
• Use command line tool nnictl
• Use NNIBoard
• Define search space
• [Use NNI sdk] - coming soon
• Config an experiment
• [Use annotation]- coming soon
• Debug

Tutorials

• How to run an experiment on local (with multiple GPUs)?
• How to run an experiment on multiple machines?
• How to run an experiment on OpenPAI?
• [Try different tuners and assessors] - coming soon
• [How to run an experiment on K8S services?] - coming soon
• [Implement a customized tuner] - coming soon
• [Implement a customized assessor] - coming soon
• [Implement a custmoized weight sharing algorithm] - coming soon
• [How to integrate NNI with your own custmoized training service] - coming soon

Best practice

• [Compare different AutoML algorithms] - coming soon
• [Serve NNI as a capability of a ML Platform] - coming soon

Contribute

This project welcomes contributions and suggestions, we use GitHub issues for tracking requests and bugs.

Issues with the good first issue label are simple and easy-to-start ones that we recommend new contributors to start with.

To set up environment for NNI development, refer to the instruction: Set up NNI developer environment

Before start coding, review and get familiar with the NNI Code Contribution Guideline: Contributing

We are in construction of the instruction for How to Debug, you are also welcome to contribute questions or suggestions on this area.

License

The entire codebase is under MIT license