Intro
This week we are going to look into the inner working of a Neural Network. As a starter we will go through the different steps involved while training a NN. Then we will play together with a ‘toy’ version of a NN built by Daniel Shiffman
Resources
- NN - Intro videos (again)
- NN - ToyNN Source
- NN - ToyNN Forked Source with shape classifier
- NN - ToyNN Example Color Predictor
- NN - Visualisation NOC examples - chapter 10
- NN - How to build a NN (XOR example)
Weekly task
Experiment with the toy NN:
- Create a shape detector
- Make the shape detector work with your webcam
- Visualise the training process
- Build a classification model using different datasets (try mnist fashion)
- Try to do some of the above or to replicate the examples provided with ToyNN using another tool and compare the performances
-
More ideas
