For the School we will be using Python for the code examples. The packages needed are:
- Matplotlib
- PyTorch
- scikit-learn
- TensorFlow
- torchdiffeq (Optional, see instructions in the package webpage)
and their installation is very easy using Anaconda Python Distribution. See instructions below:
- Anaconda Installation
- Global Anaconda installation
- Alternative installation in a new conda environment
Please bring all the software pre-installed to the School. The download size is ~1GB and we will collapse the WiFi connection if everybody tries to install it on arrival.
If you are experienced using Anaconda you can use the following ML-School-environment.yml environment file.
Anaconda Installation
By far the easiest way to have everything installed is using Anaconda Python Distribution. Anaconda installation itself is a very easy process: just follow the installation instructions for your platform in case you don’t have it already installed.
Once you have Anaconda installed, open a terminal (or an Anaconda Prompt
from the Start Menu if you are using Windows) and install the required packages using the conda
installer as described below.
Global Anaconda installation
Here we will install all the sofware so that it can be used everytime you use Anaconda Python in this computer. See below for installation into a conda environent.
PyTorch
conda install pytorch-cpu torchvision-cpu -c pytorch
For MacOs you may have to remove the -cpu
part. Check the PyTorch official installation instructions
Matplotlib, scikit-learn, TensorFlow
conda install ipython matplotlib scikit-learn tensorflow
Where we also install IPython for convenience.
For more information about TensorFlow installation (eg. GPU enabled versions) see here
Alternative installation in a new conda environment
Alternatively, we can install eveything inside a new conda environment using the provided ML-School-environment.yml environment file. This will create a new environment named ML-School
and install all the requiered software within it. Just download the file and:
conda env create -f ML-School-environment.yml
This will take a while depending on your computer and internet connection speed.
Once the environment is created you can activate it with:
conda activate ML-School
and deactivate it with:
conda deactivate
Should you want to remove it, type:
conda remove --name ML-School --all
Of course all this can also be done through the Anaconda Navigator
, should you be more accustomed to it.