Inspiration: Basic ML design, I just learned ML algorithms recently, and I use this project to test my knowledge

What it does: It recognizes hand-written digits from a test set. Accuracy is about 93%!

How I built it: Using sklearn with multiplayer perceptron, and uses numpy & matplotlib to visualize image and predictions

Challenges I ran into: Building up a proper MLP classifier, and visualizing image by reshaping it to matrix form

Accomplishments that I'm proud of: The result :-) It achieved 93% accuracy with a max_iter = 200. Also I got familiar with Kedro haha:)

What I learned: Kedro is a useful tool. It modularises your code.

What's next for Best of Kedro: Try to do more CV work. Maybe hand-pick some features and start a project from scratch

Built With

Share this project:

Updates