Kaggle Competition: cassava leaf disease detection

Introduction

This repo contains the code to complete Kaggle Competition: Cassava Leaf Disease Detection and consists of below parts:

1. Data Viwer

This is a simple interactive data viewer created using streamlit To view the dataset, you need to

  1. download dataset and unzip to ./cassava-leaf-disease-classification-data
  2. Run streamlit run data_viewer_streamlit.py and visit http://localhost:8501 and there you go.

2. Model training and predicting.

Using ResNet101V2 NN with one dense layer (256 units) as the classification model.