README.md

Herbapedia - Machine Learning Documentation

About

This repository contains workbooks for building a CNN Model that will be deployed on the Herbapedia Cloud API. For more information regarding Herbapedia, head to this link.

Tools and Python Modules Used

• Google Colab (Python Notebooks)
• Google Drive (Dataset Storage)
• Numpy v1.21.6
• TensorFlow v2.8.2
• TensorFlow.js v3.18.0
• Keras v2.8.0
• PIL v7.1.2

Training Results

Model achived 91% validation accuracy with a validation loss value of 0.2806.

Testing Trained Model

Model tested on self taken pictures, with accurate results.