tcanbolat / foodidentify

Food 101 Classification Model

FoodIDentify

Food Image Classification Model
Built with TensorFlow
Model has been trained on a data set of 101 different food types

Data set source: https://www.kaggle.com/datasets/kmader/food41?resource=download-directory

Table of Contents

1. Development Setup
2. Model Architecture
3. Training Graph
4. Setup and train model

Development Setup

• python -m venv ./venv or python3 -m venv ./venv

• source venv/bin/activate

• pip install -r requirements.txt

• python app.py

Model Architecture

Training Graph

Setup and train model

• Download data set from above link
• Copy images folder into meta folder
• source venv/bin/activate
• python setup.py ( this will create train and test folders using the images folder )
• python process.py ( this will start training the model, model will be created at the end of training )
• use plot_model_history function in common.py to plot a graph of your loss and accuracy