• MindInsight
• • Introduction
  • Installation
  • • System Environment Information Confirmation
    • Installation Methods
    • • Installation by pip
      • Installation by Source Code
      • • Downloading Source Code from Gitee
        • Compiling MindInsight
    • Installation Verification
  • Quick Start
  • Docs
  • Community
  • • Governance
    • Communication
  • Vulkan Vision
  • Contributing
  • Release Notes
  • License

MindInsight

• Introduction
• Installation
  • System Environment Information Confirmation
  • Installation Methods
    • Installation by pip
    • Installation by Source Code
      • Downloading Source Code from Gitee
      • Compiling MindInsight
  • Installation Verification
• Quick Start
• Docs
• Community
  • Governance
  • Communication
• Vulkan Vision
• Contributing
• Release Notes
• License

简体中文

Introduction

MindInsight provides MindSpore with easy-to-use debugging and tuning capabilities. During the training, data such as scalar, tensor, image, computational graph, model hyper parameter and training’s execution time can be recorded in the file for viewing and analysis through the visual page of MindInsight.

Click to view the MindInsight design document, learn more about the design.
Click to view the Tutorial documentation learn more about the MindInsight tutorial.

Installation

System Environment Information Confirmation

• The hardware platform is Ascend or GPU.
• Confirm that Python 3.7.5 is installed.
• The versions of MindInsight and MindSpore must be consistent.
• If you use source code to compile and install, the following dependencies also need to be installed:
  • Confirm that CMake 3.14.1 or later is installed.
  • Confirm that GCC 7.3.0 is installed.
  • Confirm that node.js 10.19.0 or later is installed.
  • Confirm that wheel 0.32.0 or later is installed.
  • Confirm that pybind11 2.4.3 or later is installed.
• All other dependencies are included in requirements.txt.

Installation Methods

You can install MindInsight either by pip or by source code.

Installation by pip

Install from PyPI:

pip install mindinsight

Install with customized version:

pip install https://ms-release.obs.cn-north-4.myhuaweicloud.com/{version}/MindInsight/ascend/{system}/mindinsight-{version}-cp37-cp37m-linux_{arch}.whl --trusted-host ms-release.obs.cn-north-4.myhuaweicloud.com -i https://pypi.tuna.tsinghua.edu.cn/simple

• When the network is connected, dependency items are automatically downloaded during .whl package installation. (For details about other dependency items, see requirements.txt). In other cases, you need to manually install dependency items.
• {version} denotes the version of MindInsight. For example, when you are downloading MindSpore 1.0.1, {version} should be 1.0.1.
• {arch} denotes the system architecture. For example, the Linux system you are using is x86 architecture 64-bit, {arch} should be x86_64. If the system is ARM architecture 64-bit, then it should be aarch64.
• {system} denotes the system version. For example, if you are using EulerOS ARM architecture, {system} should be euleros_aarch64. Currently, the following systems are supported by Ascend: euleros_aarch64/centos_aarch64/centos_x86/ubuntu_aarch64/ubuntu_x86. ubuntu_x86 is supported by GPU.

Installation by Source Code

Downloading Source Code from Gitee

git clone https://gitee.com/mindspore/mindinsight.git

Compiling MindInsight

You can choose any of the following installation methods:

1. Run the following command in the root directory of the source code:

   cd mindinsight
   pip install -r requirements.txt -i https://pypi.tuna.tsinghua.edu.cn/simple
   python setup.py install
   

2. Build the whl package for installation.

   Enter the root directory of the source code, first execute the MindInsight compilation script in the build directory, and then execute the command to install the whl package generated in the output directory.

   cd mindinsight
   bash build/build.sh
   pip install output/mindinsight-{version}-cp37-cp37m-linux_{arch}.whl -i https://pypi.tuna.tsinghua.edu.cn/simple
   

Installation Verification

Execute the following command:

mindinsight start

If it prompts the following information, the installation is successful:

Web address: http://127.0.0.1:8080
service start state: success

Quick Start

Before using MindInsight, the data in the training process should be recorded. When starting MindInsight, the directory of the saved data should be specified. After successful startup, the data can be viewed through the web page. Here is a brief introduction to recording training data, as well as starting and stopping MindInsight.

SummaryCollector is the interface MindSpore provides for a quick and easy collection of common data about computational graphs, loss values, learning rates, parameter weights, and so on. Below is an example of using SummaryCollector for data collection, specifying the directory where the data is stored in ./summary_dir.

...

from mindspore.train.callback import SummaryCollector
summary_collector = SummaryCollector(summary_dir='./summary_dir')
model.train(epoch=1, ds_train, callbacks=[summary_collector])

For more ways to record visual data, see the MindInsight Tutorial.

After you've collected the data, when you launch MindInsight, specify the directory in which the data has been stored.

mindinsight start --summary-base-dir ./summary_dir

After successful startup, visit http://127.0.0.1:8080 through the browser to view the web page.

Command of stopping the MindInsight service:

mindinsight stop

Docs

More details about installation guide, tutorials and APIs, please see the
User Documentation.

Community

Governance

Check out how MindSpore Open Governance works.

Communication

• MindSpore Slack - Communication platform for developers.
• IRC channel at #mindspore (only for meeting minutes logging purpose)
• Video Conferencing: TBD
• Mailing-list: https://mailweb.mindspore.cn/postorius/lists

Vulkan Vision

Vulkan Vision(V-Vision) provides an unprecedented level of detail into the execution of Vulkan applications through dynamic instrumentation. V-Vision supports analyzing AI workloads implemented using the a compute pipeline as well as traditional raster and ray-tracing Vulkan applications. To use V-Vision please refer to the build instructions. For more information, please refer to the paper published at CGO 2021.

Contributing

Welcome contributions. See our Contributor Wiki for more details.

Release Notes

The release notes, see our RELEASE.

License

Apache License 2.0