cheml

Chemical engineering ML

Overview

cheml is a toolkit for building, training, and deploying machine learning models tailored for chemical engineering applications. It provides utilities for data preprocessing, model selection, evaluation, and integration with chemical process data.

Features

• Data loaders and preprocessors for chemical datasets
• Ready-to-use ML model templates (regression, classification)
• Model evaluation and visualization tools
• Support for scikit-learn, PyTorch, and TensorFlow
• Example workflows for common chemical engineering problems

Installation

git clone https://github.com/mv-per/cheml.git
cd cheml

Environment Setup

CheML uses conda for development. Please refer to conda documentation for conda installation. You can also use a conda alternative, such as the mamba package, that is a C++ implementation of conda.

With conda setup, you can generate the environment by

chmod +x ./scripts/create_dev_env.sh
./scripts/create_dev_env.sh

then you can activate the environment by invoking:

conda activate cheml

The package uses pre-commit to lint and check typing on the files. Please initiate pre-commit on your machine by invoking:

pre-commit install

With the activated environment, one can build the package using the following command (that uses the invoke package):

inv build

Example Applications

• Predicting reaction yields
• Process optimization
• Property estimation

Contributing

Contributions are welcome! Please open issues or submit pull requests.

License

MIT License