To create a fresh conda environment with the packages you need and start jupyter:

for windows:

conda env create --file .\requirements_win.yml
conda activate ml-for-chemists-tf2
jupyter notebook

for linux:

conda env create --file requirements.yml
conda activate ml-for-chemists-tf2
jupyter notebook

This will install the GPU accelerated version of tensorflow.

If you don't have a compatible GPU, then edit the appropriate .yml file in a text editor and change tensorflow-gpu to tensorflow.

This module contains materials to teach machine learning and regression to chemistry students using chemistry relevant examples. No previous python experience is expected. A list of contents for each notebook follows, note that the 'a' notebooks are workbooks for use by the students and the 'b' notebooks contain example answers:

Teaches basic python and basic linear regression in python (2d line fitting).

Teaching multivariate regression models and model refinement whilst also getting students to redo the data analysis for a paper on using green solvents for cellulose dissolution in organic electrolyte solutions.

Teaches principal coordinate analysis and clustering algorithms (k-means) and applies these techniques to differentiating wine from different vinyards based on the wine's analytes.

Does not contain any chemistry examples, teaches scoring, dataset size and how ML algorithms can be fooled using pictorial input.

Introduces MoleculeNet and DeepChem. Teaches students how to build and use supervised regression and classification models for chemical problems like determining drug solubility or predicting action against HIV.

There will be a paper forthcoming on the application of these tutorials to teach synthetic students digital chemistry.