Machine Learning - 2024

Description

This project implements machine-learning algorithms. For more details, please refer to ./pdf/consignes.pdf.

Autoencoder for numbers with mnist_train.csv dataset for training

Requirements

Make sure you have Python >= 3.10 installed.

Install the Project

1. Clone the repository:

   git clone https://github.com/Ant0in/Projet5-IA.git

2. Navigate to the project directory:

   cd Projet5-IA/

3. Install the project dependencies (Highly recommend using a venv):

   pip install -r '.\requirements.txt'

Usage

When installed, run the project through the main.py file. The main file is located at Projet5-IA/main.py.

Required Parameters (positional):

• The dataset path dataset_path which will be used to train the model. Value must be a string and a valid model.

Training Parameters (positional, have default values)

For this section, please refer to ./pdf/consignes.pdf to understand every parameters. Default values are accessibles with the argparser in the main.py file.

• Learning Rate learning_rate. Value must be a float between 0 and 1.
• Encoded Dimension encoded_dim. Value must be a positive integer (greater than 0).
• Batch size batch_size. Value must be a positive integer (greater than 0).
• Epochs epochs. Value must be a positive integer.

Optional parameters :

• --verbose: Enables or disables verbose mode. Enabled by default.
• --log_path: Enables logging into a log file @ log_path. If no path is given, this will be disabled.

Running Examples:

python .\main.py --dataset .\dataset\mnist_train.csv --learning_rate 0.01 --batch_size 32 --epochs 10

License

This project is licensed under the MIT License. You are free to use, modify, and distribute this software.

Acknowledgements

This project was developed for the Artificial Intelligence course INFO-F311. Special thanks to Tom Lenaerts & Pascal Tribel (ULB) for their guidance and support.