This repository contains preprocessed datasets from the LeMat-Bulk dataset prepared for MatterGen-style retraining.
🚀 Loading the Dataset
Since this repository is public, you can load the dataset subsets directly without authentication using the Hugging Face datasets library:
from datasets import load_dataset
