forked from microsoft/CNTK
-
Notifications
You must be signed in to change notification settings - Fork 1
Test Configurations
Mark Hillebrand edited this page Jun 10, 2016
·
3 revisions
Here are the current configurations that are used in CNTK team environment for automated building and testing. The presented set of product versions is not restrictive, i.e. CNTK may well work in many other configurations, however this is exactly how we test our builds. Also each pull request reviewed is being tried in the same configurations and with the same set of tests as our regular builds.
- Windows 8.1 Pro (64 bit)
- Windows Server 2012 R2 Standard
- Visual Studio Ultimate 2013 Update 5
- Microsoft MPI v. 7.0
- AMD ACML 5.3.1 with Intel Fortran
- Intel® Math Kernel Library 11.3 Update 3
- NVIDIA CUDA 7.5
- NVIDIA cuDNN 4.0
- NVIDIA CUB 1.4.1
- OpenCV v.3.0.0
- zlib v.1.2.8
- libzip v.1.1.2
- Ubuntu 14.04 LTS (64 bit)
- GNU C++ 4.8.4
- Open MPI v. 1.10.2
- AMD ACML 5.3.1 with Intel Fortran
- Intel® Math Kernel Library 11.3 Update 3
- NVIDIA CUDA 7.5
- NVIDIA cuDNN 4.0
- NVIDIA CUB 1.4.1
- OpenCV v.3.0.0
- zlib v.1.2.8
- libzip v.1.1.2
Getting started
- Home
- Setup CNTK on your machine
- Tutorial
- Tutorial II
- CNTK usage overview
- Examples
- Presentations
- Multiple GPUs and machines
Configuring CNTK
- Config file overview
- Simple Network Builder
-
BrainScript Network Builder
- also see Describing Networks below
- SGD block
- Reader block
- Train, Test, Eval
- Top-level configurations
Describing Networks with BrainScript
Data readers
- Text Format Reader
- HTKMLF Reader
- LM sequence reader
- LU sequence reader
- Image reader
- Deserializers and Transforms
Evaluating CNTK Models
- Overview
- C++ Evaluation Interface
- C# Evaluation Interface
- Evaluating Hidden Layers
- C# Image Transforms for Evaluation
Advanced topics
Licenses