Image-to-Text Retrieval on FashionX (test)
99.7AccuracyGemini-2.5-Flash
Evaluation Results
| Method | Links | |
|---|---|---|
| Gemini-2.5-FlashSize=-2026.03 | 99.7 | |
| Qwen2.5-VLSize=7b2026.03 | 99.5 | |
| OmniFashionSize=3b2026.03 | 99.5 | |
| LLaVA-OneVision-1.5Size=8b2026.03 | 99.2 | |
| LLaVA-OneVision-1.5Size=4b2026.03 | 98.7 | |
| Claude-4.5-SonnetSize=-2026.03 | 98.5 | |
| Phi-4-multimodalSize=5b2026.03 | 96.7 | |
| Grok-4-FastSize=-2026.03 | 90.4 | |
| Qwen2.5-VLSize=3b2026.03 | 81.4 | |
| MiniCPM-V 2.6Size=8b2026.03 | 59.5 |