Seoul National University

A LoRA adapter on ProstT5 that predicts per-residue probability distributions over Foldseek 3Di tokens, capturing sequence-encoded conformational flexibility from MD trajectories.

A publicly available vision-language model that interprets chest X-rays and generates radiology reports, built on a CXR-specific image encoder and LLaMA-2 (non-commercial license).

Transformer-based model for predicting gene expression from histone modifications, incorporating 3D chromatin interaction data and large genomic windows to capture distal regulatory effects.