PLIP

Stanford

Open Source

Peer-reviewed

Model Information

Architecture: ViT-B/32
Parameters: 88M
Training Method: CLIP
Training Data: 200K+ image-text pairs
Release Date: 2023-07

Benchmark Rankings

HKUST PathBench: 19/21

Stanford PathBench: 19/22

STAMP: 13/13

THUNDER: 22/23

Plismbench: 12/16