Edit

News of 2016

  • Article

This are archived News of 2016.
See the actual news at this page.

December 2016

2016-12-22. V 2.0 Beta 7 Release
Highlights of this Release:

See more in the Release Notes
Get the Release from the CNTK Releases page

2016-12-13. V 2.0 Beta 6 Release
Highlights of this Release:

See more in the Release Notes
Get the Release from the CNTK Releases page

November 2016

2016-11-25. V 2.0 Beta 5 Release
Highlights of this Release:

  • The Windows binary packages are now created using the NVIDIA CUDA 8 toolkit, see the release notes for details. The CNTK-Linux binary packages are still built with CUDA 7.5. The Linux support for Cuda8 will follow shortly!
  • Performance enhancements for evaluation of bitmap images through the new EvaluateRgbImage function in the managed Eval API.
  • A new version of the CNTK NuGet package is available.
  • Stability Improvements and bug fixes, i.e. decreased memory footprint in CNTK Text Format deserializer.
  • We continue to improve documentation and tutorials on an ongoing basis, in this release we added a Sequence-to-Sequence tutorial.

See more in the Release Notes
Get the Release from the CNTK Releases page

2016-11-21. V 2.0 Beta 4 Release
Highlights of this Release:

  • New ASGD/Hogwild! training using Microsoft’s Parameter Server (Project Multiverso)
  • Distributed Scenarios now supported in CNTK Python API
  • New Memory Compression mode to reduce memory usage on GPU
  • CNTK Docker image with 1bit-SGD support
  • Stability Improvements and bug fixes

See more in the Release Notes
Get the Release from the CNTK Releases page

2016-11-11. V 2.0 Beta 3 Release
Highlights of this Release:

  • Integration with NVIDIA NCCL. Works with Linux when building CNTK from sources. See how to enable
  • The first V.2.0 Prerelease NuGet Package for CNTK Evaluation library
  • Stability Improvements and bug fixes

See more in the Release Notes
Get the Release from the CNTK Releases page

2016-11-03. V 2.0 Beta 2 Release
Highlights of this Release:

See more in the Release Notes
Get the Release from the CNTK Releases page

October 2016

2016-10-25. New CNTK Name, new Web Site and V 2.0 Beta 1 Release

CNTK becomes The Microsoft Cognitive Toolkit. See more at our new Web Site.

With the today's Release we start delivering CNTK V2 - a major upgrade of Microsoft Cognitive Toolkit.

Expect a set of Beta Releases in the Coming Weeks.

Highlights of this Release:

  • CNTK can now be used as a library with brand new C++ and Python APIs
  • New Python Examples and Tutorials
  • Support of Protocol Buffers serialization
  • Support of Fast R-CNN algorithm
  • New automated installation procedures
  • Improvements in CNTK Evaluation library including support of CNTK APIs

See more in the Release Notes. You will find there links to the materials about the new features.
Get the Release from the CNTK Releases page

2016-10-03. V 1.7.2 Binary release
This is a Hot Fix Release. It affects all users of Model Evaluation Library

If you are NOT using Model Evaluation Library you may skip this release.
If you ARE using Model Evaluation Library we strongly recommend installing version 1.7.2 instead of any previous version you might be using.

See Release Notes for details.

September 2016

2016-09-28. V 1.7.1 Binary release
Highlights of this Release:

  • Two Breaking Changes related to Layers library default initialization and fsAdagrad gradient-normalization scheme
  • Improvements in BrainScript
  • Enabling of Deterministic Algorithm enforcement
  • Improvements in Model Evaluation including the support of Evaluation for Azure Applications
  • Different Performance improvements
  • Multiple bug fixes

See more in the Release Notes (including the full list of bugs fixed)
Get the Release from the CNTK Releases page

August 2016

2016-08-31. V 1.7 Binary release
Highlights of this Release:

  • Improvements in BrainScript (new library of predefined common layer types; common random-initialization types; GRUs)
  • Support of NVIDIA cuDNN 5.1 and cuDNN RNN
  • Improvements in readers and deserializers
  • Additions to Evaluator library (Eval Client Sample, strong name for EvalWrapper)
  • New unit tests incl. Linux support
  • Python API Preview (since V.1.5)
  • Multiple bug fixes

See more in the Release Notes
Get the Release from the CNTK Releases page

2016-08-29. Two new Tutorials are available:
Image recognition (CIFAR-10) and Language understanding (ATIS).

2016-08-10. We have significantly simplified handling of Gated Recurrent Units (GRU). Read more in the corresponding article.

July 2016

2016-07-15. V 1.6 Binary release
Highlights of this Release:

  • Deconvolution and Unpooling operations
  • MKL Support (improved CPU performance)
  • CNTK Model Evaluation runtime library for Windows and Linux
  • OpenCV 3.1 (more efficient image preprocessing)
  • Simple Python support
  • Various bug fixes & optimizations

See more in the Release Notes
Get the Release from the CNTK Releases page

2016-07-12. We have further expanded Licensing options for CNTK 1bit-SGD and related components. 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.

June 2016

2016-06-23. New License Terms for CNTK 1bit-SGD and related components. The new terms provide more flexibility and enable new usage scenarios, especially in commercial environments. 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.

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.

May 2016

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

April 2016

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. CUDA 7.5 support for Windows Build: Windows project files have been updated to automatically utilize CUDA 7.5 if present

March 2016

2016-03-24. New Text Reader (CNTKTextFormatReader) is available
Read description here (./BrainScript-CNTKTextFormat-Reader.md)

February 2016

2016-02-29. Added ZIP files support to the ImageReader
Examples: (./BrainScript-Image-reader.md)
Updated build steps at (./Setup-CNTK-on-your-machine.md)

2016-02-29. Added documentation for the ImageReader ./Image-reader

2016-02-17. CNTK Contribution Guidelines are published
Check it out here: (./Contributing-to-CNTK.md)

2016-02-15. The first part of CNTK tutorial is published.
You can find it here: (./Tutorial.md)

2016-02-10. Binary release
Most up-to-date binaries are on the CNTK Releases page

January 2016

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 https://cntk.ai/ also became available.