Published in the Proceedings of the 1st Workshop on Language + Molecules (L+M 2024), at ACL 2024, Bangkok, Thailand.
The paper retrains Mistral-7B, Llama-2-7B, Llama-3-8B and Gemma-7B on 42,000 human protein sequences to generate new protein sequences. Despite the small dataset, the models performed comparably to protein-specific models trained on millions of sequences, measured by pLDDT, RMSD, TM-score and REU.
Authors: Kamyar Zeinalipour, Neda Jamshidi, Monica Bianchini, Marco Maggini, Marco Gori.
Read the paper | Code | Models and data
All the lab's language releases are listed in Language models and datasets.