Build A Dog/Cat Detector In C# With CNTK

I presented this webinar in 2019 for the students in my machine learning bootcamp.

In the webinar you’ll see me develop an object detector that uses a deep neural network to automatically categorize images of dogs and cats. I will build the neural network from scratch using the Microsoft Cognitive Toolkit.

1 ¾ Hours

English

Intermediate

Webinar Contents

Introducing Machine Learning

Current trends in AI

Jobs replaced by robots

Machine learning algorithms

The convolutional neural network

Building neural networks

The sandwich model

Convolution layers

Pooling layers

Dropout layers

Dense classification layers

The Dog & Cat Dataset

The folder structure

The dataset labels

Generating the mapping file

The C# Code

The CNTK library

The Pensar library

Building a neural network in C#

Adding convolution, pooling, dropout and dense network layers

Shaping the tensors

The C# Code (continued)

The Adam learner

The training loop

The testing loop

Demo: training the network

The Results

Overfitting 

Data augmentation

Training and validation accuracy

Saving the trained model

Categorizing single images

Watch The Webinar

Ready For More?

CHECK OUT MY AI TRAINING COURSES

Did I pique your interest? Then check out all of my AI training courses and get ready to supercharge your machine learning career today. See anything that you like?

Download The Source Code

Would you like to experiment with your own code? Just click the button to download the exact same source code that I use in the webinar. Use my code as a starting point to develop your own C# computer vision applications!

Testimonials

what my students are saying about ME