/BNet

Training Neural nets with Approximate Bayes

Primary LanguageJupyter Notebook

BNet

This project is an algorithm for training neural networks using Bayesian Linear Regression, among other things. The algorithm and demonstrations for regression and classification are in main.ipynb , plotting the results from the classification demonstration (MNIST) is in plot_mnist_results.ipynb.

The algorithm is outlined in this post. Results are found in this post.