zxcyfr/Human-Protein-Language-Model-0.5
Updated • 1
ESM-2 650M checkpoints fine-tuned on human Pfam families and clans at 0.5 and 0.9 sequence-identity thresholds for the DS-UFT benchmark.
Note Pfam-level models; sequence-identity threshold: 0.5
Note Pfam-level models; sequence-identity threshold: 0.9
Note Clan-level models; sequence-identity threshold: 0.5
Note Clan-level models; sequence-identity threshold: 0.9
Note Combined cross-family model trained on pooled Pfam-family sequences at a 0.5 sequence-identity threshold.