nouredinekn/morocco-food-ai-model
Type de record:
model
Créateur:
nou
Hôte:
# Morocco Food Classification Model By nouredine_kn
This project uses FastAI to create a machine learning model that classifies various types of Moroccan food based on images. The model is designed to recognize popular dishes and provide predictions with a probability score.
## Table of Contents
- [Project Overview]
- [Installation]
- [Usage]
- [Model Training]
- [Model Prediction]
- [Saving the Model]
## Project Overview
The model is trained on images of Moroccan food, leveraging the FastAI library for efficient data loading, model training, and predictions. It can classify dishes such as **Bastila**, **Couscous**, **Tagine**, and many more.
## Installation
To set up this project, ensure you have Python and the FastAI library installed. You can install FastAI using pip:
```bash
pip install fastai
```
Make sure you have the required dependencies for image processing:
```bash
pip install pillow
```
## Usage
1. **Data Preparation**: Place your food images in the `./morocco-food/` directory, organizing them in subdirectories according to their respective classes.
2. **Run the Notebook**: Open the Jupyter notebook (or your preferred Python environment) and execute the code cells to train and test the model.
### Model Training
The model is trained using a CNN (ResNet34 architecture) with a training-validation split of 80%-20%. You can fine-tune the model by adjusting the training parameters in the notebook:
```python
learn = cnn_learner(dls, resnet34, metrics=error_rate)
learn.fine_tune(1)
```
### Model Prediction
To test the model's prediction capability, provide an image path and execute the following code:
```python
img='path_to_your_image.jpg'
name,_,probs=learn.predict(img)
if max(probs).item()>0.7:
print(f'Food type: {name}')
print(f'Probability it\'s: {max(probs).item():.6f}')
else:
print('Not found')
```
### Saving the Model
Once you're satisfied with the training, you can save the model for future use:
```
learn.export(r'.\morocco-food-ml-by-nour …