Prerequisites

Installs

Before we begin reading about at all things language, there are a few packages that will be needed throughout the module. As with the previous two modules, it is strongly recommended that you set up a new conda environment with Python 3.7 or higher. You can create a new conda environment with many of the needed packaged by running

conda create -n week3 python=3.8 jupyter notebook numpy matplotlib numba scikit-learn nltk=3.6.5

conda install -n week3 -c conda-forge python-graphviz

We will need PyTorch as well. If on Windows or Linux, run

conda install -n week3 pytorch torchvision cpuonly -c pytorch

If on MacOS, run

conda install -n week3 pytorch torchvision -c pytorch

Make sure to activate this conda environment by running

conda activate week3

Once the new environment is activated, install

• MyGrad, MyNN, Noggin, and Gensim, by running:

pip install mygrad mynn noggin gensim cogworks-data

If you choose not to create a new conda environment, make sure that the following packages are properly installed:

• jupyter

• notebook

• numpy

• matplotlib

• numba

• scikit-learn

• nltk, make sure to install v.3.6.5, as newer versions introduce performance issues.

• python-graphviz, which should be installed via the conda-forge channel

• pytorch, where specific installation instructions for your machine can be found here

• mygrad, which can be installed via pip

• mynn, which can be installed via pip

• noggin, which can be installed via pip

• gensim, which must be installed via pip