scikit-bio/

TMVec-2s

TM-Vec 2s: A distilled Bi-LSTM for remote homology detection. 2s was trained using a teacher-student architecture.

TM-Vec is a family of protein language models fine-tuned for structure prediction, using tm-score as its predictor.

To learn more, read the original TM-Vec paper here: Protein remote homology detection and structural alignment using deep learning