Interactive MNIST Classifier

Python (PyTorch, OpenCV, Matplotlib, Streamlit) · Mar 2024 - May 2024

An app that lets users interact with multiple machine learning models for hand-written digit recognition.

  • Built and trained a multi-layer perceptron classifying hand-written digits with PyTorch

  • Devised a clean UI for model interaction with Streamlit, with input preprocess done with OpenCV

  • Collected model performance metrics like loss and confusion matrix, and visualized them with Matplotlib

Source Code

Previous
Previous

SUICA

Next
Next

Unnamed File System