Files and splits for pre-training OneProt

The .csv and .txt files correspond to the respective train, validation and test splits for different modalities used to pre-train OneProt. *seqstruc.csv structure graph modality *saprot.txt structure token modality *text.csv text modality *pocket.csv pocket modality test_all.csv a combined dataset used for the final alignment evaluation pocket_100_residues.h5 correspond to actual data for pocket modality used for pre-training OneProt Source code https://github.com/klemens-floege/oneprot/tree/main Huggingface https://huggingface.co/HelmholtzAI-FZJ/oneprot ArXiv https://arxiv.org/abs/2411.04863