Referring Object Classification on CXR8 (test)
33.5AccuracyOLIVE-R
Evaluation results
| Method | Links | |
|---|---|---|
| OLIVE-RMethod Type=Classification, Pre-Training Data=None, Backbone=Llama 2, Evaluation Protocol=Zero-shot2024.06 | 33.5 | |
| BioMedCLIPMethod Type=Classification, Pre-Training Data=PMC-15, Evaluation Protocol=Zero-shot2024.06 | 32.5 | |
| OLIVE-RGMethod Type=Generative, Pre-Training Data=COCO, Backbone=Llama 2, Evaluation Protocol=Zero-shot2024.06 | 31.2 | |
| BioMedCLIPMethod Type=Classification, Pre-Training Data=PMC-15, Evaluation Protocol=Zero-shot, Input Transformation=crop2024.06 | 23.3 | |
| CLIPMethod Type=Classification, Pre-Training Data=CLIP400M, Evaluation Protocol=Zero-shot2024.06 | 14 | |
| Random GuessMethod Type=Classification, Pre-Training Data=None, Evaluation Protocol=Zero-shot2024.06 | 12.5 | |
| Flamingo-9BMethod Type=Generative, Pre-Training Data=C4 + LAION-2b, Evaluation Protocol=8-shot2024.06 | 12.5 | |
| CLIPMethod Type=Classification, Pre-Training Data=CLIP400M, Evaluation Protocol=Zero-shot, Input Transformation=crop2024.06 | 11.2 | |
| OLIVE-GMethod Type=Generative, Pre-Training Data=COCO, Backbone=Llama 2, Evaluation Protocol=Zero-shot2024.06 | 0 |