To complete the assignments, you need a working Python installation with Jupyter and the following libraries: numpy, scipy, matplotlib, scikit-learn, pytorch, torchvision and pillow. We recommend using Miniconda although this is not mandatory.
You can find Miniconda installation instructions in the official Miniconda documentation.
Once Miniconda is installed, create a virtual environment dedicated to the homeworks using the following command
conda create --name deep_learning
You can activate the environment using the command
conda activate deep_learning
and deactivate it using
conda deactivate
To install Pytorch, go to this webpage and select the different boxes according to your system.
Be sure to activate your environment with conda activate deep_learning before entering the install command.
We recommend installing the stable version.
If you have an NVIDIA GPU on your computer pick the version of CUDA installed on your machine, otherwise pick CPU.
AMD GPUs cannot be used for deep learning as the libraries usually rely on CUDA, if you have an AMD GPU, pick the CPU version.
You can install all the remaining librairies using the following command
conda install jupyter numpy matplotlib pillow scikit-learn
or find the installation commands in the documentation: jupyter, numpy, matplotlib, scikit-learn and pillow
To open the first homework in a notebook, run the following command
jupyter notebook homework1.ipynb
A link will be prompted in the terminal, open this link in a browser and you can start editing from there.