/mnist_from_scratch

Weekend (WIP) project to learn about machine learning theory

Primary LanguageJupyter NotebookMIT LicenseMIT

mnist_from_scratch

Weekend (WIP) project to learn about machine learning theory

Basic backpropagation neural network with several features such as

  • ReLu, Sigmoid, and Softmax
  • Mean Squared Error and Cross Entropy Loss functions
  • Convolution layer (that still needs optimization as of writing this)

Other:

  • XOR Solver with Sigmoid + Softmax
  • MNIST implementation in PyTorch aswell