Skip to content

GitLab

"llm/git@developer.sourcefind.cn:OpenDAS/ollama.git" did not exist on "4fcbf1cde6ee9aa783a10d437e4b18ea030e1cf5"
Unverified Commit 5b0034e4 authored by SparkSnail's avatar SparkSnail Committed by GitHub
Browse files

Merge pull request #204 from microsoft/master 

merge master
parents 704b50e2 19173aa4
Hide whitespace changes
Inline Side-by-side
Showing with 278 additions and 125 deletions
Makefile
View file @ 5b0034e4
...@@ -54,7 +54,7 @@ NNI_NODE_FOLDER = $(NNI_DEPENDENCY_FOLDER)/nni-node-$(OS_SPEC)-x64 ...@@ -54,7 +54,7 @@ NNI_NODE_FOLDER = $(NNI_DEPENDENCY_FOLDER)/nni-node-$(OS_SPEC)-x64
NNI_NODE ?= $(BIN_FOLDER)/node NNI_NODE ?= $(BIN_FOLDER)/node
NNI_YARN_TARBALL ?= $(NNI_DEPENDENCY_FOLDER)/nni-yarn.tar.gz NNI_YARN_TARBALL ?= $(NNI_DEPENDENCY_FOLDER)/nni-yarn.tar.gz
NNI_YARN_FOLDER ?= $(NNI_DEPENDENCY_FOLDER)/nni-yarn NNI_YARN_FOLDER ?= $(NNI_DEPENDENCY_FOLDER)/nni-yarn
NNI_YARN := PATH=$(BIN_FOLDER):$${PATH} $(NNI_YARN_FOLDER)/bin/yarn NNI_YARN ?= PATH=$(BIN_FOLDER):$${PATH} $(NNI_YARN_FOLDER)/bin/yarn
## Version number ## Version number
NNI_VERSION_VALUE = $(shell git describe --tags) NNI_VERSION_VALUE = $(shell git describe --tags)
......
README.md
View file @ 5b0034e4
...@@ -27,8 +27,10 @@ The tool dispatches and runs trial jobs generated by tuning algorithms to search ...@@ -27,8 +27,10 @@ The tool dispatches and runs trial jobs generated by tuning algorithms to search
<table> <table>
<tbody> <tbody>
<tr align="center" valign="bottom"> <tr align="center" valign="bottom">
<td>
</td>
<td> <td>
<b>Supported Frameworks</b> <b>Frameworks & Libraries</b>
<img src="docs/img/bar.png"/> <img src="docs/img/bar.png"/>
</td> </td>
<td> <td>
...@@ -42,26 +44,52 @@ The tool dispatches and runs trial jobs generated by tuning algorithms to search ...@@ -42,26 +44,52 @@ The tool dispatches and runs trial jobs generated by tuning algorithms to search
</tr> </tr>
</tr> </tr>
<tr valign="top"> <tr valign="top">
<td align="center" valign="middle">
<b>Built-in</b>
</td>
<td> <td>
<ul><li><b>Supported Frameworks</b></li>
<ul> <ul>
<li>PyTorch</li> <li>PyTorch</li>
<li>TensorFlow</li>
<li>Keras</li> <li>Keras</li>
<li>TensorFlow</li>
<li>MXNet</li> <li>MXNet</li>
<li>Caffe2</li> <li>Caffe2</li>
<li>CNTK (Python language)</li> <a href="docs/en_US/SupportedFramework_Library.md">More...</a><br/>
<li>Chainer</li> </ul>
<li>Theano</li> </ul>
<ul>
<li><b>Supported Libraries</b></li>
<ul>
<li>Scikit-learn</li>
<li>XGBoost</li>
<li>LightGBM</li>
<a href="docs/en_US/SupportedFramework_Library.md">More...</a><br/>
</ul>
</ul>
<ul>
<li><b>Examples</b></li>
<ul>
<li><a href="examples/trials/mnist-distributed-pytorch">MNIST-pytorch</li></a>
<li><a href="examples/trials/mnist-distributed">MNIST-tensorflow</li></a>
<li><a href="examples/trials/mnist-keras">MNIST-keras</li></a>
<li><a href="docs/en_US/TrialExample/GbdtExample.md">Auto-gbdt</a></li>
<li><a href="docs/en_US/TrialExample/Cifar10Examples.md">Cifar10-pytorch</li></a>
<li><a href="docs/en_US/TrialExample/SklearnExamples.md">Scikit-learn</a></li>
<a href="docs/en_US/SupportedFramework_Library.md">More...</a><br/>
</ul>
</ul> </ul>
</td> </td>
<td align="left"> <td align="left" >
<a href="docs/en_US/Tuner/BuiltinTuner.md">Tuner</a> <a href="docs/en_US/Tuner/BuiltinTuner.md">Tuner</a>
<br />
<ul> <ul>
<b style="margin-left:-20px">General Tuner</b> <li><b>General Tuner</b></li>
<ul>
<li><a href="docs/en_US/Tuner/BuiltinTuner.md#Random">Random Search</a></li> <li><a href="docs/en_US/Tuner/BuiltinTuner.md#Random">Random Search</a></li>
<li><a href="docs/en_US/Tuner/BuiltinTuner.md#Evolution">Naïve Evolution</a></li> <li><a href="docs/en_US/Tuner/BuiltinTuner.md#Evolution">Naïve Evolution</a></li>
<b style="margin-left:-20px">Tuner for HPO</b> </ul>
<li><b>Tuner for <a href="docs/en_US/CommunitySharings/HpoComparision.md">HPO</a></b></li>
<ul>
<li><a href="docs/en_US/Tuner/BuiltinTuner.md#TPE">TPE</a></li> <li><a href="docs/en_US/Tuner/BuiltinTuner.md#TPE">TPE</a></li>
<li><a href="docs/en_US/Tuner/BuiltinTuner.md#Anneal">Anneal</a></li> <li><a href="docs/en_US/Tuner/BuiltinTuner.md#Anneal">Anneal</a></li>
<li><a href="docs/en_US/Tuner/BuiltinTuner.md#SMAC">SMAC</a></li> <li><a href="docs/en_US/Tuner/BuiltinTuner.md#SMAC">SMAC</a></li>
...@@ -71,14 +99,19 @@ The tool dispatches and runs trial jobs generated by tuning algorithms to search ...@@ -71,14 +99,19 @@ The tool dispatches and runs trial jobs generated by tuning algorithms to search
<li><a href="docs/en_US/Tuner/BuiltinTuner.md#MetisTuner">Metis Tuner</a></li> <li><a href="docs/en_US/Tuner/BuiltinTuner.md#MetisTuner">Metis Tuner</a></li>
<li><a href="docs/en_US/Tuner/BuiltinTuner.md#BOHB">BOHB</a></li> <li><a href="docs/en_US/Tuner/BuiltinTuner.md#BOHB">BOHB</a></li>
<li><a href="docs/en_US/Tuner/BuiltinTuner.md#GPTuner">GP Tuner</a></li> <li><a href="docs/en_US/Tuner/BuiltinTuner.md#GPTuner">GP Tuner</a></li>
<b style="margin-left:-20px">Tuner for NAS</b> </ul>
<li><b>Tuner for <a href="docs/en_US/CommunitySharings/NasComparision.md">NAS</a></b></li>
<ul>
<li><a href="docs/en_US/Tuner/BuiltinTuner.md#NetworkMorphism">Network Morphism</a></li> <li><a href="docs/en_US/Tuner/BuiltinTuner.md#NetworkMorphism">Network Morphism</a></li>
<li><a href="examples/tuners/enas_nni/README.md">ENAS</a></li> <li><a href="examples/tuners/enas_nni/README.md">ENAS</a></li>
</ul> </ul>
</ul>
<a href="docs/en_US/Assessor/BuiltinAssessor.md">Assessor</a> <a href="docs/en_US/Assessor/BuiltinAssessor.md">Assessor</a>
<ul> <ul>
<ul>
<li><a href="docs/en_US/Assessor/BuiltinAssessor.md#Medianstop">Median Stop</a></li> <li><a href="docs/en_US/Assessor/BuiltinAssessor.md#Medianstop">Median Stop</a></li>
<li><a href="docs/en_US/Assessor/BuiltinAssessor.md#Curvefitting">Curve Fitting</a></li> <li><a href="docs/en_US/Assessor/BuiltinAssessor.md#Curvefitting">Curve Fitting</a></li>
</ul>
</ul> </ul>
</td> </td>
<td> <td>
...@@ -93,11 +126,40 @@ The tool dispatches and runs trial jobs generated by tuning algorithms to search ...@@ -93,11 +126,40 @@ The tool dispatches and runs trial jobs generated by tuning algorithms to search
</ul> </ul>
</td> </td>
</tr> </tr>
<tr align="center" valign="bottom">
</td>
</tr>
<tr valign="top">
<td valign="middle">
<b>References</b>
</td>
<td style="border-top:#FF0000 solid 0px;">
<ul>
<li><a href="docs/en_US/sdk_reference.rst">Python API</a></li>
<li><a href="docs/en_US/Tutorial/AnnotationSpec.md">NNI Annotation</a></li>
<li><a href="docs/en_US/Tutorial/Installation.md">Supported OS</a></li>
</ul>
</td>
<td style="border-top:#FF0000 solid 0px;">
<ul>
<li><a href="docs/en_US/Tuner/CustomizeTuner.md">CustomizeTuner</a></li>
<li><a href="docs/en_US/Assessor/CustomizeAssessor.md">CustomizeAssessor</a></li>
</ul>
</td>
<td style="border-top:#FF0000 solid 0px;">
<ul>
<li><a href="docs/en_US/TrainingService/SupportTrainingService.md">Support TrainingService</li>
<li><a href="docs/en_US/TrainingService/HowToImplementTrainingService.md">Implement TrainingService</a></li>
</ul>
</td>
</tr>
</tbody> </tbody>
</table> </table>
## **Who should consider using NNI** ## **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 try different AutoML algorithms in their training code (model) at their local machine.
...@@ -235,60 +297,69 @@ Maybe you want to read: ...@@ -235,60 +297,69 @@ Maybe you want to read:
* [NNI overview](docs/en_US/Overview.md) * [NNI overview](docs/en_US/Overview.md)
* [Quick start](docs/en_US/Tutorial/QuickStart.md) * [Quick start](docs/en_US/Tutorial/QuickStart.md)
* [Contributing](docs/en_US/Tutorial/Contributing.md)
* [Examples](docs/en_US/examples.rst)
* [References](docs/en_US/reference.rst)
* [WebUI tutorial](docs/en_US/Tutorial/WebUI.md) * [WebUI tutorial](docs/en_US/Tutorial/WebUI.md)
* [Contributing](docs/en_US/Tutorial/Contributing.md)
## **How to** ## **How to**
* [Install NNI](docs/en_US/Tutorial/Installation.md) * [Install NNI](docs/en_US/Tutorial/Installation.md)
* [Use command line tool nnictl](docs/en_US/Tutorial/Nnictl.md) * [Use command line tool nnictl](docs/en_US/Tutorial/Nnictl.md)
* [Use NNIBoard](docs/en_US/Tutorial/WebUI.md) * [Define a trial](docs/en_US/TrialExample/Trials.md)
* [How to define search space](docs/en_US/Tutorial/SearchSpaceSpec.md)
* [How to define a trial](docs/en_US/TrialExample/Trials.md)
* [How to choose tuner/search-algorithm](docs/en_US/Tuner/BuiltinTuner.md)
* [Config an experiment](docs/en_US/Tutorial/ExperimentConfig.md) * [Config an experiment](docs/en_US/Tutorial/ExperimentConfig.md)
* [How to use annotation](docs/en_US/TrialExample/Trials.md#nni-python-annotation) * [Define search space](docs/en_US/Tutorial/SearchSpaceSpec.md)
* [choose tuner/search-algorithm](docs/en_US/Tuner/BuiltinTuner.md)
* [Use annotation](docs/en_US/TrialExample/Trials.md#nni-python-annotation)
* [Use NNIBoard](docs/en_US/Tutorial/WebUI.md)
## **Tutorials** ## **Tutorials**
* [Run an experiment on local (with multiple GPUs)](docs/en_US/TrainingService/LocalMode.md)
* [Run an experiment on OpenPAI](docs/en_US/TrainingService/PaiMode.md) * [Run an experiment on OpenPAI](docs/en_US/TrainingService/PaiMode.md)
* [Run an experiment on Kubeflow](docs/en_US/TrainingService/KubeflowMode.md) * [Run an experiment on Kubeflow](docs/en_US/TrainingService/KubeflowMode.md)
* [Run an experiment on local (with multiple GPUs)](docs/en_US/TrainingService/LocalMode.md)
* [Run an experiment on multiple machines](docs/en_US/TrainingService/RemoteMachineMode.md) * [Run an experiment on multiple machines](docs/en_US/TrainingService/RemoteMachineMode.md)
* [Try different tuners](docs/en_US/Tuner/BuiltinTuner.md) * [Try different tuners](docs/en_US/Tuner/BuiltinTuner.md)
* [Try different assessors](docs/en_US/Assessor/BuiltinAssessor.md) * [Try different assessors](docs/en_US/Assessor/BuiltinAssessor.md)
* [Implement a customized tuner](docs/en_US/Tuner/CustomizeTuner.md) * [Implement a customized tuner](docs/en_US/Tuner/CustomizeTuner.md)
* [Implement a customized assessor](docs/en_US/Assessor/CustomizeAssessor.md) * [Implement a customized assessor](docs/en_US/Assessor/CustomizeAssessor.md)
* [Implement TrainingService in NNI](docs/en_US/TrainingService/HowToImplementTrainingService.md)
* [Use Genetic Algorithm to find good model architectures for Reading Comprehension task](docs/en_US/TrialExample/SquadEvolutionExamples.md) * [Use Genetic Algorithm to find good model architectures for Reading Comprehension task](docs/en_US/TrialExample/SquadEvolutionExamples.md)
* [Advanced Neural Architecture Search](docs/en_US/AdvancedFeature/AdvancedNas.md)
## **Contribute** ## **Contribute**
This project welcomes contributions and there are many ways in which you can participate in the project, for example: This project welcomes contributions and there are many ways in which you can participate in the project, for example:
* Review [source code changes](https://github.com/microsoft/nni/pulls) * Open [bug reports](https://github.com/microsoft/nni/issues/new/choose).
* Review the [documentation](https://github.com/microsoft/nni/tree/master/docs) and make pull requests for anything from typos to new content * Request a [new feature](https://github.com/microsoft/nni/issues/new/choose).
* Suggest or ask some questions on the [How to Debug](docs/en_US/Tutorial/HowToDebug.md) guidance document.
* Find the issues tagged with ['good first issue'](https://github.com/Microsoft/nni/issues?q=is%3Aissue+is%3Aopen+label%3A%22good+first+issue%22) or ['help-wanted'](https://github.com/microsoft/nni/issues?q=is%3Aopen+is%3Aissue+label%3A%22help+wanted%22), these are simple and easy to start , we recommend new contributors to start with. * Find the issues tagged with ['good first issue'](https://github.com/Microsoft/nni/issues?q=is%3Aissue+is%3Aopen+label%3A%22good+first+issue%22) or ['help-wanted'](https://github.com/microsoft/nni/issues?q=is%3Aopen+is%3Aissue+label%3A%22help+wanted%22), these are simple and easy to start , we recommend new contributors to start with.
Before providing your hacks, there are a few simple guidelines that you need to follow:
Before providing your hacks, you can review the [Contributing Instruction](docs/en_US/Tutorial/Contributing.md) to get more information. In addition, we also provide you with the following documents:
* [NNI developer environment installation tutorial](docs/en_US/Tutorial/SetupNniDeveloperEnvironment.md)
* [How to debug](docs/en_US/Tutorial/HowToDebug.md) * [How to debug](docs/en_US/Tutorial/HowToDebug.md)
* [Code Styles & Naming Conventions](docs/en_US/Tutorial/Contributing.md) * [Customize Your Own Advisor](docs/en_US/Tuner/CustomizeAdvisor.md)
* How to Set up [NNI developer environment](docs/en_US/Tutorial/SetupNniDeveloperEnvironment.md) * [Customize Your Own Tuner](docs/en_US/Tuner/CustomizeTuner.md)
* Review the [Contributing Instruction](docs/en_US/Tutorial/Contributing.md) and get familiar with the NNI Code Contribution Guideline * [Implement customized TrainingService](docs/en_US/TrainingService/HowToImplementTrainingService.md)
## **External Repositories** ## **External Repositories**
Now we have some external usage examples run in NNI from our contributors. Thanks our lovely contributors. And welcome more and more people to join us! Now we have some external usage examples run in NNI from our contributors. Thanks our lovely contributors. And welcome more and more people to join us!
* Run [ENAS](examples/tuners/enas_nni/README.md) in NNI * Run [ENAS](examples/tuners/enas_nni/README.md) in NNI
* Run [Neural Network Architecture Search](examples/trials/nas_cifar10/README.md) in NNI * Run [Neural Network Architecture Search](examples/trials/nas_cifar10/README.md) in NNI
* [Automatic Feature Engineering](examples/trials/auto-feature-engineering/README.md) in NNI
## **Feedback** ## **Feedback**
* Open [bug reports](https://github.com/microsoft/nni/issues/new/choose).<br/>
* Request a [new feature](https://github.com/microsoft/nni/issues/new/choose).
* Discuss on the NNI [Gitter](https://gitter.im/Microsoft/nni?utm_source=badge&utm_medium=badge&utm_campaign=pr-badge&utm_content=badge) in NNI * Discuss on the NNI [Gitter](https://gitter.im/Microsoft/nni?utm_source=badge&utm_medium=badge&utm_campaign=pr-badge&utm_content=badge) in NNI
* Ask a question with NNI tags on [Stack Overflow](https://stackoverflow.com/questions/tagged/nni?sort=Newest&edited=true)or [file an issue](https://github.com/microsoft/nni/issues/new/choose)on GitHub. * Ask a question with NNI tags on [Stack Overflow](https://stackoverflow.com/questions/tagged/nni?sort=Newest&edited=true)
* We are in construction of the instruction for [How to Debug](docs/en_US/Tutorial/HowToDebug.md), you are also welcome to contribute questions or suggestions on this area. * [File an issue](https://github.com/microsoft/nni/issues/new/choose) on GitHub.
## **License** ## **License**
The entire codebase is under [MIT license](LICENSE) The entire codebase is under [MIT license](LICENSE)
README_zh_CN.md
View file @ 5b0034e4
...@@ -4,7 +4,7 @@ ...@@ -4,7 +4,7 @@
* * * * * *
[![MIT 许可证](https://img.shields.io/badge/license-MIT-brightgreen.svg)](LICENSE) [![生成状态](https://msrasrg.visualstudio.com/NNIOpenSource/_apis/build/status/Microsoft.nni)](https://msrasrg.visualstudio.com/NNIOpenSource/_build/latest?definitionId=6) [![问题](https://img.shields.io/github/issues-raw/Microsoft/nni.svg)](https://github.com/Microsoft/nni/issues?q=is%3Aissue+is%3Aopen) [![Bug](https://img.shields.io/github/issues/Microsoft/nni/bug.svg)](https://github.com/Microsoft/nni/issues?q=is%3Aissue+is%3Aopen+label%3Abug) [![拉取请求](https://img.shields.io/github/issues-pr-raw/Microsoft/nni.svg)](https://github.com/Microsoft/nni/pulls?q=is%3Apr+is%3Aopen) [![版本](https://img.shields.io/github/release/Microsoft/nni.svg)](https://github.com/Microsoft/nni/releases) [![进入 https://gitter.im/Microsoft/nni 聊天室提问](https://badges.gitter.im/Microsoft/nni.svg)](https://gitter.im/Microsoft/nni?utm_source=badge&utm_medium=badge&utm_campaign=pr-badge&utm_content=badge) [![文档状态](https://readthedocs.org/projects/nni/badge/?version=latest)](https://nni.readthedocs.io/en/latest/?badge=latest) [![MIT 许可证](https://img.shields.io/badge/license-MIT-brightgreen.svg)](LICENSE) [![生成状态](https://msrasrg.visualstudio.com/NNIOpenSource/_apis/build/status/Microsoft.nni)](https://msrasrg.visualstudio.com/NNIOpenSource/_build/latest?definitionId=6) [![问题](https://img.shields.io/github/issues-raw/Microsoft/nni.svg)](https://github.com/Microsoft/nni/issues?q=is%3Aissue+is%3Aopen) [![Bug](https://img.shields.io/github/issues/Microsoft/nni/bug.svg)](https://github.com/Microsoft/nni/issues?q=is%3Aissue+is%3Aopen+label%3Abug) [![拉取请求](https://img.shields.io/github/issues-pr-raw/Microsoft/nni.svg)](https://github.com/Microsoft/nni/pulls?q=is%3Apr+is%3Aopen) [![版本](https://img.shields.io/github/release/Microsoft/nni.svg)](https://github.com/Microsoft/nni/releases) [![进入 https://gitter.im/Microsoft/nni 聊天室提问](https://badges.gitter.im/Microsoft/nni.svg)](https://gitter.im/Microsoft/nni?utm_source=badge&utm_medium=badge&utm_campaign=pr-badge&utm_content=badge) [![文档状态](https://readthedocs.org/projects/nni/badge/?version=latest)](https://nni.readthedocs.io/zh/latest/?badge=latest)
[English](README.md) [English](README.md)
...@@ -47,40 +47,40 @@ NNI (Neural Network Intelligence) 是自动机器学习（AutoML）的工具包 ...@@ -47,40 +47,40 @@ NNI (Neural Network Intelligence) 是自动机器学习（AutoML）的工具包
</ul> </ul>
</td> </td>
<td align="left"> <td align="left">
<a href="docs/en_US/Tuner/BuiltinTuner.md">Tuner（调参器）</a> <a href="docs/zh_CN/Tuner/BuiltinTuner.md">Tuner（调参器）</a>
<br /> <br />
<ul> <ul>
<b style="margin-left:-20px">通用 Tuner</b> <b style="margin-left:-20px">通用 Tuner</b>
<li><a href="docs/en_US/Tuner/BuiltinTuner.md#Random">Random Search（随机搜索）</a></li> <li><a href="docs/zh_CN/Tuner/BuiltinTuner.md#Random">Random Search（随机搜索）</a></li>
<li><a href="docs/en_US/Tuner/BuiltinTuner.md#Evolution">Naïve Evolution（进化算法）</a></li> <li><a href="docs/zh_CN/Tuner/BuiltinTuner.md#Evolution">Naïve Evolution（进化算法）</a></li>
<b style="margin-left:-20px">超参 Tuner</b> <b style="margin-left:-20px">超参 Tuner</b>
<li><a href="docs/en_US/Tuner/BuiltinTuner.md#TPE">TPE</a></li> <li><a href="docs/zh_CN/Tuner/BuiltinTuner.md#TPE">TPE</a></li>
<li><a href="docs/en_US/Tuner/BuiltinTuner.md#Anneal">Anneal（退火算法）</a></li> <li><a href="docs/zh_CN/Tuner/BuiltinTuner.md#Anneal">Anneal（退火算法）</a></li>
<li><a href="docs/en_US/Tuner/BuiltinTuner.md#SMAC">SMAC</a></li> <li><a href="docs/zh_CN/Tuner/BuiltinTuner.md#SMAC">SMAC</a></li>
<li><a href="docs/en_US/Tuner/BuiltinTuner.md#Batch">Batch（批处理）</a></li> <li><a href="docs/zh_CN/Tuner/BuiltinTuner.md#Batch">Batch（批处理）</a></li>
<li><a href="docs/en_US/Tuner/BuiltinTuner.md#GridSearch">Grid Search（遍历搜索）</a></li> <li><a href="docs/zh_CN/Tuner/BuiltinTuner.md#GridSearch">Grid Search（遍历搜索）</a></li>
<li><a href="docs/en_US/Tuner/BuiltinTuner.md#Hyperband">Hyperband</a></li> <li><a href="docs/zh_CN/Tuner/BuiltinTuner.md#Hyperband">Hyperband</a></li>
<li><a href="docs/en_US/Tuner/BuiltinTuner.md#MetisTuner">Metis Tuner</a></li> <li><a href="docs/zh_CN/Tuner/BuiltinTuner.md#MetisTuner">Metis Tuner</a></li>
<li><a href="docs/en_US/Tuner/BuiltinTuner.md#BOHB">BOHB</a></li> <li><a href="docs/zh_CN/Tuner/BuiltinTuner.md#BOHB">BOHB</a></li>
<li><a href="docs/en_US/Tuner/BuiltinTuner.md#GPTuner">GP Tuner</a></li> <li><a href="docs/zh_CN/Tuner/BuiltinTuner.md#GPTuner">GP Tuner</a></li>
<b style="margin-left:-20px">网络结构 Tuner</b> <b style="margin-left:-20px">网络结构 Tuner</b>
<li><a href="docs/en_US/Tuner/BuiltinTuner.md#NetworkMorphism">Network Morphism</a></li> <li><a href="docs/zh_CN/Tuner/BuiltinTuner.md#NetworkMorphism">Network Morphism</a></li>
<li><a href="examples/tuners/enas_nni/README.md">ENAS</a></li> <li><a href="examples/tuners/enas_nni/README.md">ENAS</a></li>
</ul> </ul>
<a href="docs/en_US/Assessor/BuiltinAssessor.md">Assessor（评估器）</a> <a href="docs/zh_CN/Assessor/BuiltinAssessor.md">Assessor（评估器）</a>
<ul> <ul>
<li><a href="docs/en_US/Assessor/BuiltinAssessor.md#Medianstop">Median Stop（中位数终止）</a></li> <li><a href="docs/zh_CN/Assessor/BuiltinAssessor.md#Medianstop">Median Stop（中位数终止）</a></li>
<li><a href="docs/en_US/Assessor/BuiltinAssessor.md#Curvefitting">Curve Fitting（曲线拟合）</a></li> <li><a href="docs/zh_CN/Assessor/BuiltinAssessor.md#Curvefitting">Curve Fitting（曲线拟合）</a></li>
</ul> </ul>
</td> </td>
<td> <td>
<ul> <ul>
<li><a href="docs/en_US/TrainingService/LocalMode.md">本机</a></li> <li><a href="docs/zh_CN/TrainingService/LocalMode.md">本机</a></li>
<li><a href="docs/en_US/TrainingService/RemoteMachineMode.md">远程计算机</a></li> <li><a href="docs/zh_CN/TrainingService/RemoteMachineMode.md">远程计算机</a></li>
<li><b>基于 Kubernetes 的平台</b></li> <li><b>基于 Kubernetes 的平台</b></li>
<ul><li><a href="docs/en_US/TrainingService/PaiMode.md">OpenPAI</a></li> <ul><li><a href="docs/zh_CN/TrainingService/PaiMode.md">OpenPAI</a></li>
<li><a href="docs/en_US/TrainingService/KubeflowMode.md">Kubeflow</a></li> <li><a href="docs/zh_CN/TrainingService/KubeflowMode.md">Kubeflow</a></li>
<li><a href="docs/en_US/TrainingService/FrameworkControllerMode.md">基于 Kubernetes（AKS 等）的 FrameworkController</a></li> <li><a href="docs/zh_CN/TrainingService/FrameworkControllerMode.md">基于 Kubernetes（AKS 等）的 FrameworkController</a></li>
</ul> </ul>
</ul> </ul>
</td> </td>
...@@ -226,34 +226,34 @@ You can use these commands to get more information about the experiment ...@@ -226,34 +226,34 @@ You can use these commands to get more information about the experiment
点击阅读： 点击阅读：
* [NNI 概述](docs/zh_CN/Overview.md) * [NNI 概述](docs/zh_CN/Overview.md)
* [快速入门](docs/en_US/Tutorial/QuickStart.md) * [快速入门](docs/zh_CN/Tutorial/QuickStart.md)
* [贡献](docs/en_US/Tutorial/Contributing.md) * [贡献](docs/zh_CN/Tutorial/Contributing.md)
* [示例](docs/en_US/examples.rst) * [示例](docs/zh_CN/examples.rst)
* [参考](docs/en_US/reference.rst) * [参考](docs/zh_CN/reference.rst)
* [Web 界面教程](docs/en_US/Tutorial/WebUI.md) * [Web 界面教程](docs/zh_CN/Tutorial/WebUI.md)
## **入门** ## **入门**
* [安装 NNI](docs/en_US/Tutorial/Installation.md) * [安装 NNI](docs/zh_CN/Tutorial/Installation.md)
* [使用命令行工具 nnictl](docs/en_US/Tutorial/Nnictl.md) * [使用命令行工具 nnictl](docs/zh_CN/Tutorial/Nnictl.md)
* [使用 NNIBoard](docs/en_US/Tutorial/WebUI.md) * [使用 NNIBoard](docs/zh_CN/Tutorial/WebUI.md)
* [如何定义搜索空间](docs/en_US/Tutorial/SearchSpaceSpec.md) * [如何定义搜索空间](docs/zh_CN/Tutorial/SearchSpaceSpec.md)
* [如何实现 Trial 代码](docs/en_US/TrialExample/Trials.md) * [如何实现 Trial 代码](docs/zh_CN/TrialExample/Trials.md)
* [如何选择 Tuner、搜索算法](docs/en_US/Tuner/BuiltinTuner.md) * [如何选择 Tuner、搜索算法](docs/zh_CN/Tuner/BuiltinTuner.md)
* [配置 Experiment](docs/en_US/Tutorial/ExperimentConfig.md) * [配置 Experiment](docs/zh_CN/Tutorial/ExperimentConfig.md)
* [如何使用 Annotation](docs/en_US/TrialExample/Trials.md#nni-python-annotation) * [如何使用 Annotation](docs/zh_CN/TrialExample/Trials.md#nni-python-annotation)
## **教程** ## **教程**
* [在 OpenPAI 上运行 Experiment](docs/en_US/TrainingService/PaiMode.md) * [在 OpenPAI 上运行 Experiment](docs/zh_CN/TrainingService/PaiMode.md)
* [在 Kubeflow 上运行 Experiment](docs/en_US/TrainingService/KubeflowMode.md) * [在 Kubeflow 上运行 Experiment](docs/zh_CN/TrainingService/KubeflowMode.md)
* [在本机运行 Experiment (支持多 GPU 卡)](docs/en_US/TrainingService/LocalMode.md) * [在本机运行 Experiment (支持多 GPU 卡)](docs/zh_CN/TrainingService/LocalMode.md)
* [在多机上运行 Experiment](docs/en_US/TrainingService/RemoteMachineMode.md) * [在多机上运行 Experiment](docs/zh_CN/TrainingService/RemoteMachineMode.md)
* [尝试不同的 Tuner](docs/en_US/Tuner/BuiltinTuner.md) * [尝试不同的 Tuner](docs/zh_CN/Tuner/BuiltinTuner.md)
* [尝试不同的 Assessor](docs/en_US/Assessor/BuiltinAssessor.md) * [尝试不同的 Assessor](docs/zh_CN/Assessor/BuiltinAssessor.md)
* [实现自定义 Tuner](docs/en_US/Tuner/CustomizeTuner.md) * [实现自定义 Tuner](docs/zh_CN/Tuner/CustomizeTuner.md)
* [实现自定义 Assessor](docs/en_US/Assessor/CustomizeAssessor.md) * [实现自定义 Assessor](docs/zh_CN/Assessor/CustomizeAssessor.md)
* [使用进化算法为阅读理解任务找到好模型](docs/en_US/TrialExample/SquadEvolutionExamples.md) * [使用进化算法为阅读理解任务找到好模型](docs/zh_CN/TrialExample/SquadEvolutionExamples.md)
## **贡献** ## **贡献**
...@@ -265,10 +265,10 @@ You can use these commands to get more information about the experiment ...@@ -265,10 +265,10 @@ You can use these commands to get more information about the experiment
在提交代码前，需要遵循以下的简单准则： 在提交代码前，需要遵循以下的简单准则：
* [如何调试](docs/en_US/Tutorial/HowToDebug.md) * [如何调试](docs/zh_CN/Tutorial/HowToDebug.md)
* [代码风格和命名约定](docs/en_US/Tutorial/Contributing.md) * [代码风格和命名约定](docs/zh_CN/Tutorial/Contributing.md)
* 如何设置 [NNI 开发环境](docs/zh_CN/Tutorial/SetupNniDeveloperEnvironment.md) * 如何设置 [NNI 开发环境](docs/zh_CN/Tutorial/SetupNniDeveloperEnvironment.md)
* 查看[贡献说明](docs/en_US/Tutorial/Contributing.md)并熟悉 NNI 的代码贡献指南 * 查看[贡献说明](docs/zh_CN/Tutorial/Contributing.md)并熟悉 NNI 的代码贡献指南
## **外部代码库** ## **外部代码库**
......
deployment/docker/Dockerfile
View file @ 5b0034e4
...@@ -49,7 +49,7 @@ RUN DEBIAN_FRONTEND=noninteractive && \ ...@@ -49,7 +49,7 @@ RUN DEBIAN_FRONTEND=noninteractive && \
# #
# update pip # update pip
# #
RUN python3 -m pip install --upgrade pip RUN python3 -m pip install --upgrade pip setuptools==39.1.0
# numpy 1.14.3 scipy 1.1.0 # numpy 1.14.3 scipy 1.1.0
RUN python3 -m pip --no-cache-dir install \ RUN python3 -m pip --no-cache-dir install \
......
docs/en_US/AdvancedFeature/GeneralNasInterfaces.md
View file @ 5b0034e4
...@@ -120,7 +120,7 @@ Here, `nni.training_update` is to do some update on the full graph. In enas_mode ...@@ -120,7 +120,7 @@ Here, `nni.training_update` is to do some update on the full graph. In enas_mode
**\*oneshot_mode\***: following the training approach in [this paper][6]. Different from enas_mode which trains the full graph by training large numbers of subgraphs, in oneshot_mode the full graph is built and dropout is added to candidate inputs and also added to candidate ops' outputs. Then this full graph is trained like other DL models. [Detailed Description](#OneshotMode). (currently only supported on tensorflow). **\*oneshot_mode\***: following the training approach in [this paper][6]. Different from enas_mode which trains the full graph by training large numbers of subgraphs, in oneshot_mode the full graph is built and dropout is added to candidate inputs and also added to candidate ops' outputs. Then this full graph is trained like other DL models. [Detailed Description](#OneshotMode). (currently only supported on tensorflow).
To use oneshot_mode, you should add one more field in the `trial` config as shown below. In this mode, no need to specify tuner in the config file as it does not need tuner. (Note that you still need to specify a tuner (any tuner) in the config file for now.) Also, no need to add `nni.training_update` in this mode, because no special processing (or update) is needed during training. To use oneshot_mode, you should add one more field in the `trial` config as shown below. In this mode, though there is no need to use tuner, you still need to specify a tuner (any tuner) in the config file for now. Also, no need to add `nni.training_update` in this mode, because no special processing (or update) is needed during training.
```diff ```diff
trial: trial:
command: your command to run the trial command: your command to run the trial
...@@ -132,7 +132,7 @@ trial: ...@@ -132,7 +132,7 @@ trial:
**\*darts_mode\***: following the training approach in [this paper][3]. It is similar to oneshot_mode. There are two differences, one is that darts_mode only add architecture weights to the outputs of candidate ops, the other is that it trains model weights and architecture weights in an interleaved manner. [Detailed Description](#DartsMode). **\*darts_mode\***: following the training approach in [this paper][3]. It is similar to oneshot_mode. There are two differences, one is that darts_mode only add architecture weights to the outputs of candidate ops, the other is that it trains model weights and architecture weights in an interleaved manner. [Detailed Description](#DartsMode).
To use darts_mode, you should add one more field in the `trial` config as shown below. In this mode, also no need to specify tuner in the config file as it does not need tuner. (Note that you still need to specify a tuner (any tuner) in the config file for now.) To use darts_mode, you should add one more field in the `trial` config as shown below. In this mode, though there is no need to use tuner, you still need to specify a tuner (any tuner) in the config file for now.
```diff ```diff
trial: trial:
command: your command to run the trial command: your command to run the trial
...@@ -156,9 +156,9 @@ for _ in range(num): ...@@ -156,9 +156,9 @@ for _ in range(num):
### enas_mode ### enas_mode
In enas_mode, the compiled trial code builds the full graph (rather than subgraph), it receives a chosen architecture and training this architecture on the full graph for a mini-batch, then request another chosen architecture. It is supported by [NNI multi-phase](./multiPhase.md). In enas_mode, the compiled trial code builds the full graph (rather than subgraph), it receives a chosen architecture and training this architecture on the full graph for a mini-batch, then request another chosen architecture. It is supported by [NNI multi-phase](./MultiPhase.md).
Specifically, for trials using tensorflow, we create and use tensorflow variable as signals, and tensorflow conditional functions to control the search space (full-graph) to be more flexible, which means it can be changed into different sub-graphs (multiple times) depending on these signals. [Here]() is an example for enas_mode. Specifically, for trials using tensorflow, we create and use tensorflow variable as signals, and tensorflow conditional functions to control the search space (full-graph) to be more flexible, which means it can be changed into different sub-graphs (multiple times) depending on these signals. [Here](https://github.com/microsoft/nni/tree/master/examples/trials/mnist-nas/enas_mode) is an example for enas_mode.
<a name="OneshotMode"></a> <a name="OneshotMode"></a>
...@@ -168,7 +168,7 @@ Below is the figure to show where dropout is added to the full graph for one lay ...@@ -168,7 +168,7 @@ Below is the figure to show where dropout is added to the full graph for one lay
![](../../img/oneshot_mode.png) ![](../../img/oneshot_mode.png)
As suggested in the [paper][6], a dropout method is implemented to the inputs for every layer. The dropout rate is set to r^(1/k), where 0 < r < 1 is a hyper-parameter of the model (default to be 0.01) and k is number of optional inputs for a specific layer. The higher the fan-in, the more likely each possible input is to be dropped out. However, the probability of dropping out all optional_inputs of a layer is kept constant regardless of its fan-in. Suppose r = 0.05. If a layer has k = 2 optional_inputs then each one will independently be dropped out with probability 0.051/2 ≈ 0.22 and will be retained with probability 0.78. If a layer has k = 7 optional_inputs then each one will independently be dropped out with probability 0.051/7 ≈ 0.65 and will be retained with probability 0.35. In both cases, the probability of dropping out all of the layer's optional_inputs is 5%. The outputs of candidate ops are dropped out through the same way. [Here]() is an example for oneshot_mode. As suggested in the [paper][6], a dropout method is implemented to the inputs for every layer. The dropout rate is set to r^(1/k), where 0 < r < 1 is a hyper-parameter of the model (default to be 0.01) and k is number of optional inputs for a specific layer. The higher the fan-in, the more likely each possible input is to be dropped out. However, the probability of dropping out all optional_inputs of a layer is kept constant regardless of its fan-in. Suppose r = 0.05. If a layer has k = 2 optional_inputs then each one will independently be dropped out with probability 0.051/2 ≈ 0.22 and will be retained with probability 0.78. If a layer has k = 7 optional_inputs then each one will independently be dropped out with probability 0.051/7 ≈ 0.65 and will be retained with probability 0.35. In both cases, the probability of dropping out all of the layer's optional_inputs is 5%. The outputs of candidate ops are dropped out through the same way. [Here](https://github.com/microsoft/nni/tree/master/examples/trials/mnist-nas/oneshot_mode) is an example for oneshot_mode.
<a name="DartsMode"></a> <a name="DartsMode"></a>
...@@ -178,7 +178,7 @@ Below is the figure to show where architecture weights are added to the full gra ...@@ -178,7 +178,7 @@ Below is the figure to show where architecture weights are added to the full gra
![](../../img/darts_mode.png) ![](../../img/darts_mode.png)
In `nni.training_update`, tensorflow MomentumOptimizer is used to train the architecture weights based on the pass `loss` and `feed_dict`. [Here]() is an example for darts_mode. In `nni.training_update`, tensorflow MomentumOptimizer is used to train the architecture weights based on the pass `loss` and `feed_dict`. [Here](https://github.com/microsoft/nni/tree/master/examples/trials/mnist-nas/darts_mode) is an example for darts_mode.
### [__TODO__] Multiple trial jobs for One-Shot NAS ### [__TODO__] Multiple trial jobs for One-Shot NAS
......
docs/en_US/CommunitySharings/HpoComparision.md
View file @ 5b0034e4
# Hyperparameter Optimization Comparison # Hyper Parameter Optimization Comparison
*Posted by Anonymous Author* *Posted by Anonymous Author*
Comparison of Hyperparameter Optimization algorithms on several problems. Comparison of Hyperparameter Optimization (HPO) algorithms on several problems.
Hyperparameter Optimization algorithms are list below: Hyperparameter Optimization algorithms are list below:
......
docs/en_US/SupportedFramework_Library.md 0 → 100644
View file @ 5b0034e4
# Framework and Library Supports
With the built-in Python API, NNI naturally supports the hyper parameter tuning and neural network search for all the AI frameworks and libraries who support Python models(`version >= 3.5`). NNI had also provided a set of examples and tutorials for some of the popular scenarios to make jump start easier.
## Supported AI Frameworks
* <b>[PyTorch]</b> https://github.com/pytorch/pytorch
<ul>
<li><a href="../../examples/trials/mnist-distributed-pytorch">MNIST-pytorch</a><br/></li>
<li><a href="TrialExample/Cifar10Examples.md">CIFAR-10</a><br/></li>
<li><a href="../../examples/trials/kaggle-tgs-salt/README.md">TGS salt identification chanllenge</a><br/></li>
<li><a href="../../examples/trials/network_morphism/README.md">Network_morphism</a><br/></li>
</ul>
* <b>[TensorFlow]</b> https://github.com/tensorflow/tensorflow
<ul>
<li><a href="../../examples/trials/mnist-distributed">MNIST-tensorflow</a><br/></li>
<li><a href="../../examples/trials/ga_squad/README.md">Squad</a><br/></li>
</ul>
* <b>[Keras]</b> https://github.com/keras-team/keras
<ul>
<li><a href="../../examples/trials/mnist-keras">MNIST-keras</a><br/></li>
<li><a href="../../examples/trials/network_morphism/README.md">Network_morphism</a><br/></li>
</ul>
* <b>[MXNet]</b> https://github.com/apache/incubator-mxnet
* <b>[Caffe2]</b> https://github.com/BVLC/caffe
* <b>[CNTK (Python language)]</b> https://github.com/microsoft/CNTK
* <b>[Spark MLlib]</b> http://spark.apache.org/mllib/
* <b>[Chainer]</b> https://chainer.org/
* <b>[Theano]</b> https://pypi.org/project/Theano/ <br/>
You are encouraged to [contribute more examples](Tutorial/Contributing.md) for other NNI users.
## Supported Library
NNI also supports all libraries written in python.Here are some common libraries, including some algorithms based on GBDT: XGBoost, CatBoost and lightGBM.
* <b>[Scikit-learn]</b> https://scikit-learn.org/stable/
<ul>
<li><a href="TrialExample/SklearnExamples.md">Scikit-learn</a><br/></li>
</ul>
* <b>[XGBoost]</b> https://xgboost.readthedocs.io/en/latest/
* <b>[CatBoost]</b> https://catboost.ai/
* <b>[LightGBM]</b> https://lightgbm.readthedocs.io/en/latest/
<ul>
<li><a href="TrialExample/GbdtExample.md">Auto-gbdt</a><br/></li>
</ul>
Here is just a small list of libraries that supported by NNI. If you are interested in NNI, you can refer to the [tutorial](TrialExample/Trials.md) to complete your own hacks.
In addition to the above examples, we also welcome more and more users to apply NNI to your own work, if you have any doubts, please refer [Write a Trial Run on NNI](TrialExample/Trials.md). In particular, if you want to be a contributor of NNI, whether it is the sharing of examples , writing of Tuner or otherwise, we are all looking forward to your participation.More information please refer to [here](Tutorial/Contributing.md).
docs/en_US/TrainingService/PaiMode.md
View file @ 5b0034e4
...@@ -54,7 +54,7 @@ Compared with [LocalMode](LocalMode.md) and [RemoteMachineMode](RemoteMachineMod ...@@ -54,7 +54,7 @@ Compared with [LocalMode](LocalMode.md) and [RemoteMachineMode](RemoteMachineMod
* shmMB * shmMB
* Optional key. Set the shmMB configuration of OpenPAI, it set the shared memory for one task in the task role. * Optional key. Set the shmMB configuration of OpenPAI, it set the shared memory for one task in the task role.
* authFile * authFile
* Optional key, Set the auth file path for private registry while using PAI mode, [Refer](https://github.com/microsoft/pai/blob/2ea69b45faa018662bc164ed7733f6fdbb4c42b3/docs/faq.md#q-how-to-use-private-docker-registry-job-image-when-submitting-an-openpai-job). * Optional key, Set the auth file path for private registry while using PAI mode, [Refer](https://github.com/microsoft/pai/blob/2ea69b45faa018662bc164ed7733f6fdbb4c42b3/docs/faq.md#q-how-to-use-private-docker-registry-job-image-when-submitting-an-openpai-job), you can prepare the authFile and simply provide the local path of this file, NNI will upload this file to HDFS for you.
Once complete to fill NNI experiment config file and save (for example, save as exp_pai.yml), then run the following command Once complete to fill NNI experiment config file and save (for example, save as exp_pai.yml), then run the following command
``` ```
......
docs/en_US/TrainingService/SupportTrainingService.md 0 → 100644
View file @ 5b0034e4
# TrainingService
NNI TrainingService provides the training platform for running NNI trial jobs. NNI supports [local](./LocalMode.md), [remote](./RemoteMachineMode.md), [pai](./PaiMode.md), [kubeflow](./KubeflowMode.md) and [frameworkcontroller](./FrameworkControllerMode.md) built-in training services.
NNI not only provides few built-in training service options, but also provides a method for customers to build their own training service easily.
## Built-in TrainingService
|TrainingService|Brief Introduction|
|---|---|
|[__Local__](./LocalMode.md)|NNI supports running an experiment on local machine, called local mode. Local mode means that NNI will run the trial jobs and nniManager process in same machine, and support gpu schedule function for trial jobs.|
|[__Remote__](./RemoteMachineMode.md)|NNI supports running an experiment on multiple machines through SSH channel, called remote mode. NNI assumes that you have access to those machines, and already setup the environment for running deep learning training code. NNI will submit the trial jobs in remote machine, and schedule suitable machine with enouth gpu resource if specified.|
|[__Pai__](./PaiMode.md)|NNI supports running an experiment on [OpenPAI](https://github.com/Microsoft/pai) (aka pai), called pai mode. Before starting to use NNI pai mode, you should have an account to access an [OpenPAI](https://github.com/Microsoft/pai) cluster. See [here](https://github.com/Microsoft/pai#how-to-deploy) if you don't have any OpenPAI account and want to deploy an OpenPAI cluster. In pai mode, your trial program will run in pai's container created by Docker.|
|[__Kubeflow__](./KubeflowMode.md)|NNI supports running experiment on [Kubeflow](https://github.com/kubeflow/kubeflow), called kubeflow mode. Before starting to use NNI kubeflow mode, you should have a Kubernetes cluster, either on-premises or [Azure Kubernetes Service(AKS)](https://azure.microsoft.com/en-us/services/kubernetes-service/), a Ubuntu machine on which [kubeconfig](https://kubernetes.io/docs/concepts/configuration/organize-cluster-access-kubeconfig/) is setup to connect to your Kubernetes cluster. If you are not familiar with Kubernetes, [here](https://kubernetes.io/docs/tutorials/kubernetes-basics/) is a good start. In kubeflow mode, your trial program will run as Kubeflow job in Kubernetes cluster.|
|[__FrameworkController__](./FrameworkControllerMode.md)|NNI supports running experiment using [FrameworkController](https://github.com/Microsoft/frameworkcontroller), called frameworkcontroller mode. FrameworkController is built to orchestrate all kinds of applications on Kubernetes, you don't need to install Kubeflow for specific deep learning framework like tf-operator or pytorch-operator. Now you can use FrameworkController as the training service to run NNI experiment.|
## TrainingService Implementation
TrainingService is designed to be easily implemented, we define an abstract class TrainingService as the parent class of all kinds of TrainingService, users just need to inherit the parent class and complete their own child class if they want to implement customized TrainingService.
The abstract function in TrainingService is shown below:
```
abstract class TrainingService {
public abstract listTrialJobs(): Promise<TrialJobDetail[]>;
public abstract getTrialJob(trialJobId: string): Promise<TrialJobDetail>;
public abstract addTrialJobMetricListener(listener: (metric: TrialJobMetric) => void): void;
public abstract removeTrialJobMetricListener(listener: (metric: TrialJobMetric) => void): void;
public abstract submitTrialJob(form: JobApplicationForm): Promise<TrialJobDetail>;
public abstract updateTrialJob(trialJobId: string, form: JobApplicationForm): Promise<TrialJobDetail>;
public abstract get isMultiPhaseJobSupported(): boolean;
public abstract cancelTrialJob(trialJobId: string, isEarlyStopped?: boolean): Promise<void>;
public abstract setClusterMetadata(key: string, value: string): Promise<void>;
public abstract getClusterMetadata(key: string): Promise<string>;
public abstract cleanUp(): Promise<void>;
public abstract run(): Promise<void>;
}
```
The parent class of TrainingService has a few abstract functions, users need to inherit the parent class and implement all of these abstract functions.
For more information about how to write your own TrainingService, please [refer](https://github.com/SparkSnail/nni/blob/dev-trainingServiceDoc/docs/en_US/TrainingService/HowToImplementTrainingService.md).
docs/en_US/Tuner/SmacTuner.md
View file @ 5b0034e4
...@@ -5,4 +5,4 @@ SMAC Tuner on NNI ...@@ -5,4 +5,4 @@ SMAC Tuner on NNI
[SMAC](https://www.cs.ubc.ca/~hutter/papers/10-TR-SMAC.pdf) is based on Sequential Model-Based Optimization (SMBO). It adapts the most prominent previously used model class (Gaussian stochastic process models) and introduces the model class of random forests to SMBO, in order to handle categorical parameters. The SMAC supported by nni is a wrapper on [the SMAC3 github repo](https://github.com/automl/SMAC3). [SMAC](https://www.cs.ubc.ca/~hutter/papers/10-TR-SMAC.pdf) is based on Sequential Model-Based Optimization (SMBO). It adapts the most prominent previously used model class (Gaussian stochastic process models) and introduces the model class of random forests to SMBO, in order to handle categorical parameters. The SMAC supported by nni is a wrapper on [the SMAC3 github repo](https://github.com/automl/SMAC3).
Note that SMAC on nni only supports a subset of the types in [search space spec](../Tutorial/SearchSpaceSpec.md), including `choice`, `randint`, `uniform`, `loguniform`, `quniform(q=1)`. Note that SMAC on nni only supports a subset of the types in [search space spec](../Tutorial/SearchSpaceSpec.md), including `choice`, `randint`, `uniform`, `loguniform`, `quniform`.
\ No newline at end of file \ No newline at end of file
docs/en_US/Tutorial/Nnictl.md
View file @ 5b0034e4
...@@ -123,6 +123,7 @@ Debug mode will disable version check function in Trialkeeper. ...@@ -123,6 +123,7 @@ Debug mode will disable version check function in Trialkeeper.
|------|------|------ |------| |------|------|------ |------|
|id| False| |The id of the experiment you want to stop| |id| False| |The id of the experiment you want to stop|
|--port, -p| False| |Rest port of the experiment you want to stop| |--port, -p| False| |Rest port of the experiment you want to stop|
|--all, -a| False| |Stop all of experiments|
* Details & Examples * Details & Examples
...@@ -144,10 +145,10 @@ Debug mode will disable version check function in Trialkeeper. ...@@ -144,10 +145,10 @@ Debug mode will disable version check function in Trialkeeper.
nnictl stop --port 8080 nnictl stop --port 8080
``` ```
4. Users could use 'nnictl stop all' to stop all experiments. 4. Users could use 'nnictl stop --all' to stop all experiments.
```bash ```bash
nnictl stop all nnictl stop --all
``` ```
5. If the id ends with *, nnictl will stop all experiments whose ids matchs the regular. 5. If the id ends with *, nnictl will stop all experiments whose ids matchs the regular.
......
docs/en_US/Tutorial/SearchSpaceSpec.md
View file @ 5b0034e4
...@@ -10,11 +10,11 @@ To define a search space, users should define the name of variable, the type of ...@@ -10,11 +10,11 @@ To define a search space, users should define the name of variable, the type of
```yaml ```yaml
{ {
"dropout_rate":{"_type":"uniform","_value":[0.1,0.5]}, "dropout_rate": {"_type": "uniform", "_value": [0.1, 0.5]},
"conv_size":{"_type":"choice","_value":[2,3,5,7]}, "conv_size": {"_type": "choice", "_value": [2, 3, 5, 7]},
"hidden_size":{"_type":"choice","_value":[124, 512, 1024]}, "hidden_size": {"_type": "choice", "_value": [124, 512, 1024]},
"batch_size":{"_type":"choice","_value":[50, 250, 500]}, "batch_size": {"_type": "choice", "_value": [50, 250, 500]},
"learning_rate":{"_type":"uniform","_value":[0.0001, 0.1]} "learning_rate": {"_type": "uniform", "_value": [0.0001, 0.1]}
} }
``` ```
...@@ -25,55 +25,54 @@ Take the first line as an example. `dropout_rate` is defined as a variable whose ...@@ -25,55 +25,54 @@ Take the first line as an example. `dropout_rate` is defined as a variable whose
All types of sampling strategies and their parameter are listed here: All types of sampling strategies and their parameter are listed here:
* {"_type":"choice","_value":options} * `{"_type": "choice", "_value": options}`
* Which means the variable's value is one of the options. Here 'options' should be a list. Each element of options is a number of string. It could also be a nested sub-search-space, this sub-search-space takes effect only when the corresponding element is chosen. The variables in this sub-search-space could be seen as conditional variables. * Which means the variable's value is one of the options. Here 'options' should be a list. Each element of options is a number of string. It could also be a nested sub-search-space, this sub-search-space takes effect only when the corresponding element is chosen. The variables in this sub-search-space could be seen as conditional variables.
* An simple [example](https://github.com/microsoft/nni/tree/master/examples/trials/mnist-nested-search-space/search_space.json) of [nested] search space definition. If an element in the options list is a dict, it is a sub-search-space, and for our built-in tuners you have to add a key '_name' in this dict, which helps you to identify which element is chosen. Accordingly, here is a [sample](https://github.com/microsoft/nni/tree/master/examples/trials/mnist-nested-search-space/sample.json) which users can get from nni with nested search space definition. Tuners which support nested search space is as follows: * An simple [example](https://github.com/microsoft/nni/tree/master/examples/trials/mnist-nested-search-space/search_space.json) of [nested] search space definition. If an element in the options list is a dict, it is a sub-search-space, and for our built-in tuners you have to add a key `_name` in this dict, which helps you to identify which element is chosen. Accordingly, here is a [sample](https://github.com/microsoft/nni/tree/master/examples/trials/mnist-nested-search-space/sample.json) which users can get from nni with nested search space definition. Tuners which support nested search space is as follows:
- Random Search - Random Search
- TPE - TPE
- Anneal - Anneal
- Evolution - Evolution
* {"_type":"randint","_value":[lower, upper]} * `{"_type": "randint", "_value": [lower, upper]}`
* Choosing a random integer from `lower` (inclusive) to `upper` (exclusive).
* Note: Different tuners may interpret `randint` differently. Some (e.g., TPE, GridSearch) treat integers from lower
to upper as unordered ones, while others respect the ordering (e.g., SMAC). If you want all the tuners to respect
the ordering, please use `quniform` with `q=1`.
* For now, we implement the "randint" distribution with "quniform", which means the variable value is a value like round(uniform(lower, upper)). The type of chosen value is float. If you want to use integer value, please convert it explicitly. * `{"_type": "uniform", "_value": [low, high]}`
* {"_type":"uniform","_value":[low, high]}
* Which means the variable value is a value uniformly between low and high. * Which means the variable value is a value uniformly between low and high.
* When optimizing, this variable is constrained to a two-sided interval. * When optimizing, this variable is constrained to a two-sided interval.
* {"_type":"quniform","_value":[low, high, q]} * `{"_type": "quniform", "_value": [low, high, q]}`
* Which means the variable value is a value like clip(round(uniform(low, high) / q) * q, low, high), where the clip operation is used to constraint the generated value in the bound. For example, for _value specified as [0, 10, 2.5], possible values are [0, 2.5, 5.0, 7.5, 10.0]; For _value specified as [2, 10, 5], possible values are [2, 5, 10]. * Which means the variable value is a value like `clip(round(uniform(low, high) / q) * q, low, high)`, where the clip operation is used to constraint the generated value in the bound. For example, for `_value` specified as [0, 10, 2.5], possible values are [0, 2.5, 5.0, 7.5, 10.0]; For `_value` specified as [2, 10, 5], possible values are [2, 5, 10].
* Suitable for a discrete value with respect to which the objective is still somewhat "smooth", but which should be bounded both above and below. If you want to uniformly choose integer from a range [low, high], you can write `_value` like this: `[low, high, 1]`. * Suitable for a discrete value with respect to which the objective is still somewhat "smooth", but which should be bounded both above and below. If you want to uniformly choose integer from a range [low, high], you can write `_value` like this: `[low, high, 1]`.
* {"_type":"loguniform","_value":[low, high]} * `{"_type": "loguniform", "_value": [low, high]}`
* Which means the variable value is a value drawn from a range [low, high] according to a loguniform distribution like exp(uniform(log(low), log(high))), so that the logarithm of the return value is uniformly distributed. * Which means the variable value is a value drawn from a range [low, high] according to a loguniform distribution like exp(uniform(log(low), log(high))), so that the logarithm of the return value is uniformly distributed.
* When optimizing, this variable is constrained to be positive. * When optimizing, this variable is constrained to be positive.
* {"_type":"qloguniform","_value":[low, high, q]} * `{"_type": "qloguniform", "_value": [low, high, q]}`
* Which means the variable value is a value like clip(round(loguniform(low, high) / q) * q, low, high), where the clip operation is used to constraint the generated value in the bound. * Which means the variable value is a value like `clip(round(loguniform(low, high) / q) * q, low, high)`, where the clip operation is used to constraint the generated value in the bound.
* Suitable for a discrete variable with respect to which the objective is "smooth" and gets smoother with the size of the value, but which should be bounded both above and below. * Suitable for a discrete variable with respect to which the objective is "smooth" and gets smoother with the size of the value, but which should be bounded both above and below.
* {"_type":"normal","_value":[mu, sigma]} * `{"_type": "normal", "_value": [mu, sigma]}`
* Which means the variable value is a real value that's normally-distributed with mean mu and standard deviation sigma. When optimizing, this is an unconstrained variable. * Which means the variable value is a real value that's normally-distributed with mean mu and standard deviation sigma. When optimizing, this is an unconstrained variable.
* {"_type":"qnormal","_value":[mu, sigma, q]} * `{"_type": "qnormal", "_value": [mu, sigma, q]}`
* Which means the variable value is a value like round(normal(mu, sigma) / q) * q * Which means the variable value is a value like `round(normal(mu, sigma) / q) * q`
* Suitable for a discrete variable that probably takes a value around mu, but is fundamentally unbounded. * Suitable for a discrete variable that probably takes a value around mu, but is fundamentally unbounded.
* {"_type":"lognormal","_value":[mu, sigma]} * `{"_type": "lognormal", "_value": [mu, sigma]}`
* Which means the variable value is a value drawn according to `exp(normal(mu, sigma))` so that the logarithm of the return value is normally distributed. When optimizing, this variable is constrained to be positive.
* Which means the variable value is a value drawn according to exp(normal(mu, sigma)) so that the logarithm of the return value is normally distributed. When optimizing, this variable is constrained to be positive.
* {"_type":"qlognormal","_value":[mu, sigma, q]} * `{"_type": "qlognormal", "_value": [mu, sigma, q]}`
* Which means the variable value is a value like round(exp(normal(mu, sigma)) / q) * q * Which means the variable value is a value like `round(exp(normal(mu, sigma)) / q) * q`
* Suitable for a discrete variable with respect to which the objective is smooth and gets smoother with the size of the variable, which is bounded from one side. * Suitable for a discrete variable with respect to which the objective is smooth and gets smoother with the size of the variable, which is bounded from one side.
* {"_type":"mutable_layer","_value":{mutable_layer_infomation}} * `{"_type": "mutable_layer", "_value": {mutable_layer_infomation}}`
* Type for [Neural Architecture Search Space][1]. Value is also a dictionary, which contains key-value pairs representing respectively name and search space of each mutable_layer. * Type for [Neural Architecture Search Space][1]. Value is also a dictionary, which contains key-value pairs representing respectively name and search space of each mutable_layer.
* For now, users can only use this type of search space with annotation, which means that there is no need to define a json file for search space since it will be automatically generated according to the annotation in trial code. * For now, users can only use this type of search space with annotation, which means that there is no need to define a json file for search space since it will be automatically generated according to the annotation in trial code.
* For detailed usage, please refer to [General NAS Interfaces][1]. * For detailed usage, please refer to [General NAS Interfaces][1].
......
docs/en_US/Tutorial/WebUI.md
View file @ 5b0034e4
...@@ -11,7 +11,7 @@ Click the tab "Overview". ...@@ -11,7 +11,7 @@ Click the tab "Overview".
* If your experiment have many trials, you can change the refresh interval on here. * If your experiment have many trials, you can change the refresh interval on here.
![](../../img/webui-img/refresh-interval.png) ![](../../img/webui-img/refresh-interval.png)
* Support to review and download the experiment result and nni-manager/dispatcher log file from the download. * Support to review and download the experiment result and nni-manager/dispatcher log file from the "View" button.
![](../../img/webui-img/download.png) ![](../../img/webui-img/download.png)
* You can click the learn about in the error box to track experiment log message if the experiment's status is error. * You can click the learn about in the error box to track experiment log message if the experiment's status is error.
...@@ -50,13 +50,11 @@ Click the tab "Intermediate Result" to see the lines graph. ...@@ -50,13 +50,11 @@ Click the tab "Intermediate Result" to see the lines graph.
![](../../img/webui-img/trials_intermeidate.png) ![](../../img/webui-img/trials_intermeidate.png)
We set a filter function for the intermediate result graph because that the trials may have many intermediate results in the training progress. You need to provide data if you want to use the filter button to see the trend of some trial. The trial may have many intermediate results in the training progress. In order to see the trend of some trials more clearly, we set a filtering function for the intermediate result graph.
What data should be written in the first input? Maybe you find an intermediate count those trials became better or worse. In other word, it's an important and concerned intermediate count. Just input it into the first input. You may find that these trials will get better or worse at one of intermediate results. In other words, this is an important and relevant intermediate result. To take a closer look at the point here, you need to enter its corresponding abscissa value at #Intermediate.
After selecting the intermeidate count, you should input your focus metric's range on this intermediate count. Yes, it's the min and max value. Like this picture, I choose the intermeidate count is 9 and the metric's range is 60-80. And then input the range of metrics on this intermedia result. Like below picture, it chooses No. 4 intermediate result and set the range of metrics to 0.8-1.
As a result, I filter these trials that the metric's range is 20-60 on the 13 intermediate count.
![](../../img/webui-img/filter-intermediate.png) ![](../../img/webui-img/filter-intermediate.png)
## View trials status ## View trials status
......
docs/en_US/conf.py
View file @ 5b0034e4
...@@ -28,7 +28,7 @@ author = 'Microsoft' ...@@ -28,7 +28,7 @@ author = 'Microsoft'
# The short X.Y version # The short X.Y version
version = '' version = ''
# The full version, including alpha/beta/rc tags # The full version, including alpha/beta/rc tags
release = 'v0.7' release = 'v1.0'
# -- General configuration --------------------------------------------------- # -- General configuration ---------------------------------------------------
......
docs/img/webui-img/compare.png
View replaced file @ 704b50e2 View file @ 5b0034e4
docs/img/webui-img/compare.png

48.8 KB | W: | H:

docs/img/webui-img/compare.png

43 KB | W: | H:

docs/img/webui-img/compare.png
docs/img/webui-img/compare.png
docs/img/webui-img/compare.png
docs/img/webui-img/compare.png
  • 2-up
  • Swipe
  • Onion skin
docs/img/webui-img/download.png
View replaced file @ 704b50e2 View file @ 5b0034e4
docs/img/webui-img/download.png

18.6 KB | W: | H:

docs/img/webui-img/download.png

23.3 KB | W: | H:

docs/img/webui-img/download.png
docs/img/webui-img/download.png
docs/img/webui-img/download.png
docs/img/webui-img/download.png
  • 2-up
  • Swipe
  • Onion skin
docs/img/webui-img/over1.png
View replaced file @ 704b50e2 View file @ 5b0034e4
docs/img/webui-img/over1.png

61 KB | W: | H:

docs/img/webui-img/over1.png

63.5 KB | W: | H:

docs/img/webui-img/over1.png
docs/img/webui-img/over1.png
docs/img/webui-img/over1.png
docs/img/webui-img/over1.png
  • 2-up
  • Swipe
  • Onion skin
docs/img/webui-img/over2.png
View replaced file @ 704b50e2 View file @ 5b0034e4
docs/img/webui-img/over2.png

29.9 KB | W: | H:

docs/img/webui-img/over2.png

33.1 KB | W: | H:

docs/img/webui-img/over2.png
docs/img/webui-img/over2.png
docs/img/webui-img/over2.png
docs/img/webui-img/over2.png
  • 2-up
  • Swipe
  • Onion skin
docs/img/webui-img/refresh-interval.png
View replaced file @ 704b50e2 View file @ 5b0034e4
docs/img/webui-img/refresh-interval.png

24.2 KB | W: | H:

docs/img/webui-img/refresh-interval.png

24.5 KB | W: | H:

docs/img/webui-img/refresh-interval.png
docs/img/webui-img/refresh-interval.png
docs/img/webui-img/refresh-interval.png
docs/img/webui-img/refresh-interval.png
  • 2-up
  • Swipe
  • Onion skin
Markdown is supported
0% or .
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment