-
Notifications
You must be signed in to change notification settings - Fork 0
Home
CNTK - Computational Network Toolkit - is a unified deep-learning toolkit by Microsoft Research. This Wiki describes how to setup CNTK, how to use, configure and test it and how to define your own networks. It is our hope that the community will take advantage of CNTK to share ideas more quickly through the exchange of open source working code.
Here are a few pages to get started:
- Setting up CNTK on your machine
- Tutorials, Examples, etc..
- CNTK usage overview
- How to contribute to CNTK
Note to search the pages of this wiki, in the search box, type: Language:Markdown yourSearchText
For a general introduction to computational networks and the core algorithms, or to cite the work, please refer to the Microsoft Technical Report MSR-TR-2014-112: "An Introduction to Computational Networks and the Computational Network Toolkit", whose source files are contained in the CNTK repository in the folder Documentation/CNTK-TechReport/lyx. Please refer to the wiki for the most updated information on features and user guidelines. The book will be less frequently edited.
2016-07-15. V 1.6 Binary release
CNTK v.1.6 binaries are on the CNTK Releases page
2016-07-12. We have further expanded Licensing options for CNTK 1bit-SGD and related components. See the details at the Wiki page. These new options are an extension of the new CNTK 1bit-SGD License that we have announced on Jun 23, 2016.
2016-07-05. CNTK now supports Deconvolution and Unpooling. See the usage example in the Network number 4 in MNIST Sample.
2016-06-23. New License Terms for CNTK 1bit-SGD and related components.
Effective immediately the License Terms for CNTK 1bit-SGD and related components have changed. The new Terms provide more flexibility and enable new usage scenarios, especially in commercial environments. Read the new Terms at the standard location. Please note, that while the new Terms are significantly more flexible comparing to the previous ones, they are still more restrictive than the main CNTK License. Consequently everything described in Enabling 1bit-SGD section of the Wiki remains valid.
2016-06-20. A post on Intel MKL and CNTK is published in the Intel IT Peer Network
2016-06-16. V 1.5 Binary release. NuGet Package with CNTK Model Evaluation Libraries.
NuGet Package is added to CNTK v.1.5 binaries. See CNTK Releases page and NuGet Package description.
2016-06-15. CNTK now supports building against a custom Intel® Math Kernel Library (MKL). See setup instructions on how to set this up for your platform.
2016-06-10. See CNTK v.1.5 binary release announcement in the official Microsoft Research Blog
2016-06-08. V 1.5 Binary release
CNTK v.1.5 binaries are on the CNTK Releases page
2016-06-01. An updated version of the network-description language has been made available under the new BrainScript Network Builder, which features full expression parsing, recursive functions, and more.
2016-05-19. A 1-hour talk describing CNTK, how to use it, and how it works, has been posted at Presentations.
2016-05-16. An example illustrating Using CNTK with ResNet is added to the codebase. The example contains some pre-trained models that can be used in various applications.
2016-05-16. CNTK Wiki now has FAQ Page
2016-05-05. CNTK now supports BlockMomentum Stochastic Gradient Descent (SGD) algorithm.
See the details in the Multiple GPUs and machines Wiki section
2016-05-03. New transformations are implemented for Image Reader.
See the description in the Image Reader Wiki section
2016-04-25. V 1.1 Binary release
CNTK v.1.1 binaries are on the CNTK Releases page
2016-04-12. CNTK is available as Azure Virtual Machines and Docker Containers
2016-04-12. Added support for ND convolution and ND pooling and CPU support for cudnn
layout in convolution, pooling and batch normalization nodes.
Read documentation on convolution, pooling and batch normalization nodes.
2016-04-05. CUDA7.5 support for Windows Build: Windows project files have been updated to automatically utilize CUDA 7.5 if present
2016-03-24. New Text Reader (CNTKTextFormatReader) is available
Read description here https://github.com/Microsoft/CNTK/wiki/CNTKTextFormat-Reader
2016-02-29. Added ZIP files support to the ImageReader
Examples: https://github.com/Microsoft/CNTK/wiki/Image-reader
Updated build steps at https://github.com/Microsoft/CNTK/wiki/Setup-CNTK-on-your-machine
2016-02-29. Added documentation for the ImageReader https://github.com/Microsoft/CNTK/wiki/Image-reader
2016-02-17. CNTK Contribution Guidelines are published
Check it out here: https://github.com/Microsoft/CNTK/wiki/Contributing-to-CNTK
2016-02-15. The first part of CNTK tutorial is published.
You can find it here: https://github.com/Microsoft/CNTK/wiki/Tutorial
2016-02-10. Binary release
Most up-to-date binaries are on the CNTK Releases page
2016-01-26. First binary release after the move to GitHub is published (GPU only).
2016-01-25 13:00 GMT. CNTK new repository in GitHub went public! Together with GitHub repository a new CNTK web site http://www.cntk.ai/ also became available.
Getting Started
Additional Documentation
How to use CNTK
Using CNTK Models in Your Code
- Overview
- Nuget Package for Evaluation
- C++ Evaluation Interface
- C# Evaluation Interface
- Evaluating Hidden Layers
- C# Image Transforms for Evaluation
- C# Multi-model Evaluation
- Evaluate in Azure
Advanced topics
Licenses
Source Code & Development