Colored MNIST Challenge

The Colored MNIST Challenge involves classifying images of colored digits. The goal is to train a model that can accurately classify these digits based on the provided training and testing datasets.

Dataset

The dataset consists of colored images of digits along with their corresponding labels. The dataset files are provided in .gz format.

• colored_mnist_images.gz: Training images
• colored_mnist_targets.gz: Training labels
• colored_mnist_test_images.gz: Test images

Files

• colored_mnist_images.gz: Contains the training images in compressed format.
• colored_mnist_targets.gz: Contains the training labels in compressed format.
• colored_mnist_test_images.gz: Contains the test images in compressed format.
• Colored_MNIST_challenge.ipynb: Baseline Jupyter notebook for the challenge.

Getting Started

1. Clone the Repository

   Clone the repository to your local machine:

   git clone https://github.com/Automatants/Colored_MNIST.git

2. Install Dependencies

   If you don't want to install all the dependencies on your machine, create a virtual environment :

   python -m venv .venv

   To activate the venv :

   Windows :

   source .venv/Scripts/activate

   macOS/Linux :

   source .venv/bin/activate

   To deactivate the venv :

   deactivate

   Use the requirements.txt file to install the dependencies :

   pip install -r requirements.txt

3. Code !

   Here, the challenge is to find a relevant model and data preprocessing to obtain the best possible accuracy. In the baseline, all necessary steps - from data import to model training - are already completed.

4. Post your predictions on the CS sharing

   Post the "predictions.npy" generated by your modified baseline in the site in order to get your score. You will have to create an account and a team before. https://sharing.cs-campus.fr/compete/113

Good luck & have fun