Model Checkpoints: Spectral Reconstruction and Cross-Modal Transfer via Lorentzian Recurrence

Read full text Summarise with VIRAT
Authors: K. Tubhyam
Cited by: 0

Summary

No summary yet. VIRAT can read this paper and write a plain-language summary of its aim, methods, key findings and limitations.

Summarise with VIRAT Free to try with your Science Catalogs account.

Abstract

21 trained PyTorch model checkpoints for all experiments reported in "Spectral Reconstruction and Cross-Modal Transfer via Lorentzian Recurrence: Oscillatory State Space Models for Vibrational Spectroscopy" (K. Tubhyam, 2026). Includes: (1) E1 architecture benchmark — D-LinOSS, CNN-1D, S4D, and Transformer checkpoints across 3 seeds (12 files); (2) E1-A architecture ablation — full, no_moe, no_post_tf, no_vib, backbone_only (5 files); (3) E5 loss ablation — full, no_physics, no_contrastive, no_vib (4 files). All models trained on QMe14S (103,992 molecules, B3LYP/TZVP DFT IR spectra). See README.md inside the archive for loading instructions and per-checkpoint validation losses.

Fields of Study & Subjects

Spectroscopy and Chemometric Analyses Spectroscopy and Quantum Chemical Studies Spectroscopy Techniques in Biomedical and Chemical Research

Publication details

Journal / source
Zenodo (CERN European Organization for Nuclear Research)
Publication date
23 March 2026
Language
EN
Document type
Dataset