AmpLM: Target-conditioned antimicrobial peptide generation with measured-MIC preference learning

This package accompanies “AmpLM: Target-conditioned antimicrobial peptide generation with measured-MIC preference learning”. AmpLM (antimicrobial peptide language model) generates peptide candidates using a broad or pathogen-specific target condition. The package provides generation code, training datasets, model weights from four training stages, and the manuscript figures and tables. The main manuscript and Supplementary Information are supplied separately as Word and PDF files alongside this archive.