DIVE Lab at TAMU

Welcome to the Hugging Face organization for the DIVE Lab at Texas A&M University. We strive to seek synergies between foundational and use-inspired themes. Our foundational research centers on developing innovative models and algorithms in the fields of machine learning, geometric deep learning, language models and agents. Our use-inspired research aims at tackling challenges in various scientific and engineering disciplines, including physics-informed modeling and simulations, biology, drug discovery, quantum physics and chemistry, materials science, molecular dynamics and simulation, fluid dynamics, and partial differential equations, among others.

The datasets/benchmarks available in our Hugging Face repository are described below:

Sys2Bench

Sys2bench is a benchmark designed to evaluate Large Language Models’ reasoning and plannning abilities across arithmetic, logical, common, algorithmic reasoning and planning.

Link: https://huggingface.co/datasets/divelab/Sys2Bench

ShockCast

Supersonic flow datasets from A Two-Phase Deep Learning Framework for Adaptive Time-Stepping in High-Speed Flow Modeling. These datasets model a multiphase coal dust explosion and a circular blast.

Link: https://huggingface.co/datasets/divelab/ShockCast

PubChemQCR

PubChemQCR is a dataset that contains the DFT relaxation trajectory of ~3.5 million small molecules, which can facilitate the development of machine learning interatomic potential (MLIP) models.

Link: https://huggingface.co/datasets/divelab/PubChemQCR

OrbEvo

TDDFT dataset for time-dependent electronic wavefunction simulation from Orbital Transformers for Predicting Wavefunctions in Time-Dependent Density Functional Theory.

Link: https://huggingface.co/datasets/divelab/OrbEvo

All other scientific and engineering projects from our lab can be found at the following link:

Artificial Intelligence Research for Science (AIRS): https://github.com/divelab/AIRS/tree/main

Connect

• GitHub
• Website