Multiclass image-text classification on Galar
14.3Macro F1CapCLIP-M (ViT-L/14)
Evaluation results
| Method | Links | ||
|---|---|---|---|
| CapCLIP-M (ViT-L/14)Zero-shot=true, Backbone=ViT-L/142026.05 | 14.3 | 22.9 | |
| CapCLIP-M (ViT-B/16)Zero-shot=true, Backbone=ViT-B/162026.05 | 11 | 19.8 | |
| CapCLIP-B (ViT-L/14)Zero-shot=true, Backbone=ViT-L/142026.05 | 8.4 | 17.2 | |
| CapCLIP-MIX (ViT-B/16)Zero-shot=true, Backbone=ViT-B/162026.05 | 7.1 | 17 | |
| CapCLIP-B (ViT-B/16)Zero-shot=true, Backbone=ViT-B/162026.05 | 6.7 | 16.1 | |
| MedImageInsightZero-shot=true2026.05 | 6.5 | 5.2 | |
| CapCLIP-MIX (ViT-L/14)Zero-shot=true, Backbone=ViT-L/142026.05 | 6 | 14.5 | |
| CLIP-OpenAIZero-shot=true2026.05 | 5.1 | 6.4 | |
| MedCLIPZero-shot=true2026.05 | 4.7 | 5.7 | |
| BioMedCLIPZero-shot=true2026.05 | 4 | 3.5 |