MNIST Neural Network Challenge

Build your own neural network to classify hand-written digits with the MNIST dataset. Your task is to use what you learn in the workshop, together with the hints provided at the bottom of the Colab notebook, to create the best machine learning classification model out of the crowd.

Pose your question in the Kurius discord group and mention Richard if you're really stuck and need help!

Instructions:

  1. Open the Colab Notebook: https://drive.google.com/file/d/1kGpq3BZQE3act3Z2S2ZdNza_QkSTsqfd/view?usp=sharing
  2. Make a copy of the Notebook using your Google Drive account
  3. Rewatch the workshop video to get yourself familiar with how the Colab notebook works, and carefully read the hints at the bottom of the notebook
  4. Only make changes to the settings of the neural network (Block A) and training (Block B)
  5. Run the code and screenshot the Test Loss and Test Accuracy print-out
  6. Upload the screenshots
  7. Submit a link to the ipynb file

Related Challenges