Published May 20, 2024 | Version v1
Model Open

Aligning protein generative models with experimental fitness via Direct Preference Optimization

Authors/Creators

  • 1. ROR icon Stanford University

Description

Model weights for the ProteinDPO model trained with the Paired DPO Objective. ProteinDPO is introduced in the paper "Aligning protein generative models with experimental fitness via Direct Preference Optimization"

Files

paired_weights.pt.zip

Files (526.9 MB)

Name Size Download all
md5:720c20fcf844ff5d3877bff14821c975
526.9 MB Preview Download