Datasets for MulitModalSpectralTransformer project

Data Repository for "Advancing Structure Elucidation with a Flexible Multi-Spectral AI Model"

This repository contains the complete dataset and code required to reproduce the results presented in our publication. Contents include:

Code and Models:

Full implementation of MultiModalSpectralTransformer (MMST)

Pre-trained model weights

Dependencies: SGNN, Mol2Mol, and ChemProp-IR libraries

Training and Validation Data:

ZINC dataset (4M molecules with simulated spectra)

PubChem dataset (1.5M molecules with simulated spectra)

Simulated spectral data for experimental validation compounds

Manually peak-picked experimental NMR and IR data

Complete experimental validation results and benchmarks

 For implementation details and usage instructions, please refer to our GitHub repository.