/tools/pre-training-via-denoising

pre-training-via-denoising

shehzaidi/pre-training-via-denoising

99 stars14 forksPythonAdded February 4, 2026
summary

This repository provides an implementation of a pre-training method via denoising for predicting molecular properties using the TorchMD-NET architecture. It includes pre-training on the PCQM4Mv2 dataset and fine-tuning for specific molecular property predictions like HOMO/LUMO.

description

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

topics

denoisingdenoising-autoencodersgraph-neural-networksmolecular-property-predictionmolecular-structurespre-trainingpytorchscore-matchingdeep-learning

Ratings

N/A
0 ratings

Rate this tool:

Comments (0)

Sign in to leave a comment.

No comments yet. Be the first to share your thoughts!