shehzaidi/pre-training-via-denoising

Official implementation of pre-training via denoising for TorchMD-NET

40
/ 100
Emerging

This project helps chemists and materials scientists predict properties of new molecules by 'pre-training' a model with vast amounts of existing molecular structure data. It takes in 3D molecular structures (like from the PCQM4Mv2 dataset) and outputs a trained model capable of accurately predicting specific molecular properties, such as HOMO/LUMO energy levels. It's designed for researchers who need to develop highly accurate predictive models for drug discovery or materials science applications.

No commits in the last 6 months.

Use this if you need to predict properties of molecules with high accuracy, especially when you have a large dataset of molecular structures to leverage for initial model training.

Not ideal if you don't have access to computational resources (GPUs) or are not comfortable working with command-line interfaces to run machine learning models.

molecular-modeling drug-discovery materials-science cheminformatics quantum-chemistry
Stale 6m No Package No Dependents
Maintenance 0 / 25
Adoption 9 / 25
Maturity 16 / 25
Community 15 / 25

How are scores calculated?

Stars

99

Forks

14

Language

Python

License

MIT

Last pushed

Mar 02, 2023

Commits (30d)

0

Get this data via API

curl "https://pt-edge.onrender.com/api/v1/quality/ml-frameworks/shehzaidi/pre-training-via-denoising"

Open to everyone — 100 requests/day, no key needed. Get a free key for 1,000/day.