Direct Recommendation on Amazon Beauty 5-core filtering (leave-one-out)
56.44Hit Rate @10 (H@10)VLM2Rec
Evaluation results
| Method | Links | ||||
|---|---|---|---|---|---|
| VLM2RecModel Category=Multimodal-based, Backbone=Qwen2.5-VL-3B2026.03 | 56.44 | 68.22 | 38.24 | 41.21 | |
| LLMSFTModel Category=Text-based, Training Strategy=SFT2026.03 | 49.63 | 63.47 | 30.85 | 34.34 | |
| SASRecModel Category=ID-based2026.03 | 42.8 | 52.83 | 29.84 | 32.36 | |
| VLMSFT(Int.)Model Category=Multimodal-based, Fusion Strategy=Internal, Training Strategy=SFT2026.03 | 41 | 57.94 | 29.41 | 34.98 | |
| VLMSFT(Ext.)Model Category=Multimodal-based, Fusion Strategy=External, Training Strategy=SFT2026.03 | 40.38 | 63.72 | 27.44 | 33 | |
| GRU4RecModel Category=ID-based2026.03 | 39.03 | 48.98 | 26.24 | 28.75 | |
| LLM2RecModel Category=Text-based2026.03 | 32.82 | 47.69 | 18.54 | 22.27 | |
| LLMEMBModel Category=Text-based2026.03 | 31.09 | 58.36 | 23.88 | 28.1 | |
| SLIM+Model Category=Text-based2026.03 | 27.15 | 41.12 | 15.04 | 18.54 | |
| LLMVanillaModel Category=Text-based, Training Strategy=Vanilla2026.03 | 23 | 34.28 | 13.53 | 16.36 | |
| VLMPromptModel Category=Multimodal-based2026.03 | 21.93 | 33.67 | 11.9 | 14.85 | |
| VLMVanilla(Int.)Model Category=Multimodal-based, Fusion Strategy=Internal2026.03 | 21.46 | 29.86 | 13.58 | 15.69 | |
| CLIPModel Category=Multimodal-based2026.03 | 21.14 | 28.19 | 13.18 | 14.94 | |
| VLMVanilla(Ext.)Model Category=Multimodal-based, Fusion Strategy=External2026.03 | 20.96 | 33.79 | 10.97 | 14.19 | |
| NoteLLM-2Model Category=Multimodal-based2026.03 | 19.26 | 35.36 | 9.29 | 13.31 | |
| SLIMModel Category=Text-based2026.03 | 14.12 | 27.24 | 6.61 | 9.87 | |
| BERTModel Category=Text-based2026.03 | 13.72 | 26.02 | 6.71 | 9.78 |