Conversational Recommendation on MUSE Multimodal Fashion (test)
10.2R@1HARPO
Evaluation Results
| Method | Links | ||||||||
|---|---|---|---|---|---|---|---|---|---|
| HARPO2026.04 | 10.2 | 38.6 | 58.4 | 19.8 | 26.4 | 72 | 68 | 52.4 | |
| Qwen2-VL-7Bfine-tuning_protocol=Wang et al. (2025)2026.04 | 8.4 | 34.2 | 52.8 | 17.2 | 23.1 | 61 | 57 | 46.8 | |
| LLaVA-Next-8Bfine-tuning_protocol=Wang et al. (2025)2026.04 | 5.2 | 25.4 | 44.2 | 12 | 16.2 | 52 | 48 | 38 | |
| GPT-4V2026.04 | 4.4 | 23.2 | 42.6 | 10.8 | 14.8 | 54 | 50 | 37.4 | |
| UniCRSadaptation_type=text-only2026.04 | 1.6 | 11.8 | 27.4 | 5.1 | 7.2 | 36 | 32 | 22.9 |