Image-to-Text Retrieval on CUB-200
0.898Spearman Correlation (S)GeoFlowVLM
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| GeoFlowVLMPre-training Dataset=CC1M, Evaluation Protocol=Zero-shot2026.05 | 0.898 | 0.767 | |
| AsymVLM-PSDPre-training Dataset=CC1M, Evaluation Protocol=Zero-shot2026.05 | 0.867 | 0.612 | |
| AsymVLM-vMFPre-training Dataset=CC1M, Evaluation Protocol=Zero-shot2026.05 | 0.729 | 0.612 | |
| GeoFlowVLMPre-training Dataset=CC1M, Evaluation Protocol=Zero-shot2026.05 | 0.723 | 0.576 | |
| ProbVLMPre-training Dataset=CC1M, Evaluation Protocol=Zero-shot2026.05 | 0.636 | 0.335 | |
| ProbVLMPre-training Dataset=CC1M, Evaluation Protocol=Zero-shot2026.05 | 0.538 | 0.327 | |
| GroVEPre-training Dataset=CC1M, Evaluation Protocol=Zero-shot2026.05 | 0.152 | 0.026 | |
| AsymVLM-PSDPre-training Dataset=CC1M, Evaluation Protocol=Zero-shot2026.05 | 0.073 | 0 | |
| AsymVLM-vMFPre-training Dataset=CC1M, Evaluation Protocol=Zero-shot2026.05 | 0.073 | 0 | |
| GroVEPre-training Dataset=CC1M, Evaluation Protocol=Zero-shot2026.05 | 0.043 | 0.061 |