Image-to-Text Retrieval on Educontent zero-shot
12.1R@1 (Zero-Shot)HOMIE
Evaluation results
| Method | Links | |||
|---|---|---|---|---|
| HOMIEModel Category=MLLM-Based Models, Backbone=Qwen3-VL-8B, Evaluation Protocol=Zero-shot2025.02 | 12.1 | 29.8 | 38 | |
| HOMIEModel Category=MLLM-Based Models, Backbone=Qwen3-VL-2B, Evaluation Protocol=Zero-shot2025.02 | 11.1 | 28.5 | 36.5 | |
| BiomedCLIPModel Category=Pathology Dual-Encoder Models, Evaluation Protocol=Zero-shot2025.02 | 3.4 | 8.4 | 14.9 | |
| PubmedCLIPModel Category=Pathology Dual-Encoder Models, Evaluation Protocol=Zero-shot2025.02 | 0.2 | 1.3 | 2.6 |