swh:1:snp:f50ab94432af916b5fb8b4ad831e8dddded77084
Raw File
Tip revision: 184e8398ddc0e52a8355b083b85c5dae12e9ead8 authored by t-makrei on 31 August 2017, 14:26:32 UTC
added python routine which recognizes a transpose operation with more than 5 unflattened axes and performs an alternative transpose operation in this case.
Tip revision: 184e839
README.md
# CNTK Tuturial: Image recognition

## Overview

This hands-on lab shows how to implement convolution-based image recognition with CNTK. We will start with a common convolutional image-recognition architecture, add Batch Normalization, and then extend it into a Residual Network (ResNet-20).

## Tutorial

Please find a detailed tutorial that uses the data and configurations in this folder on out website at https://docs.microsoft.com/en-us/cognitive-toolkit/Hands-On-Labs-Image-Recognition
back to top