Raccoon Image Classifier

Summary

This is an image classifier built with Tensorflow. The model can classify whether something is a raccoon, dog, cat, or something else. This repo is packaged into a Streamlit app can be found on my blogpost here:

• https://garbage-time.github.io/posts/raccoon-classifier/

The model is still a work-in-progress as it tends to over-classify images as 'dogs', due to sample inadequacies.

Contents

• model.py fits the model. If the user wishes to train their own model, they will have to construct their own dataset. This script is a modified variant of tensorflow's image classification tutorial found here: https://tensorflow.org/tutorials/images/classification
• app.py creates the streamlit applicaton.
• raccoon.tflite is the trained model. You may load it to try out the classifier outside of the application.