Image Retrieval on Fashion200k (test)
27.1Recall@1MOON3.0
Evaluation results
| Method | Links | |||||
|---|---|---|---|---|---|---|
| MOON3.0Setting=Zero-shot2026.04 | 27.1 | 62.72 | — | — | 51.96 | |
| CoSMoRe-implementation=false, Extra Captions=false2022.03 | 23.3 | 50.4 | 69.3 | 47.7 | — | |
| VALRe-implementation=false, Extra Captions=true2022.03 | 22.9 | 50.8 | 72.7 | 48.8 | — | |
| ComposeAE2020.06 | 22.8 | 55.3 | 73.4 | — | — | |
| ARTEMISBackbone=RN18, Text Encoder=LSTM, Extra Captions=false2022.03 | 21.5 | 51.1 | 70.5 | 47.7 | — | |
| VALRe-implementation=false, Extra Captions=false2022.03 | 21.2 | 49 | 68.8 | 46.3 | — | |
| TIRGRe-implementation=true, Backbone=RN18, Text Encoder=LSTM, Extra Captions=false2022.03 | 20.8 | 50.7 | 69.9 | 47.1 | — | |
| TIRGRe-implementation=true, Backbone=RN18, Text Encoder=BiGRU, Extra Captions=false2022.03 | 20.2 | 49.4 | 69.8 | 46.5 | — | |
| ARTEMISBackbone=RN18, Text Encoder=BiGRU, Extra Captions=false2022.03 | 20.2 | 49.3 | 69.3 | 46.2 | — | |
| TIRGtext_encoder=BERT, query_type=Complete Text Query2020.06 | 19.9 | 51.7 | 71.8 | — | — | |
| JVSMRe-implementation=false, Extra Captions=true2022.03 | 19 | 52.1 | 70 | 47 | — | |
| MOON3.0Setting=Zero-shot2026.04 | 18.53 | 48.72 | — | — | 38.98 | |
| TIRGquery_type=Complete Text Query2020.06 | 18.1 | 52.4 | 73.1 | — | — | |
| LBFRe-implementation=false, Extra Captions=false2022.03 | 17.8 | 48.4 | 68.5 | 44.9 | — | |
| TIRGtext_encoder=BERT2020.06 | 14.2 | 41.9 | 63.3 | — | — | |
| Qwen3-VL-EmbeddingSetting=Zero-shot2026.04 | 14.14 | 41.98 | — | — | 31.87 | |
| TIRG2018.12 | 14.1 | 42.5 | 63.8 | — | — | |
| TIRG2020.06 | 14.1 | 42.5 | 63.8 | — | — | |
| TIRGRe-implementation=false, Extra Captions=false, Source=Vo et al. (2019)2022.03 | 14.1 | 42.5 | 63.8 | 40.1 | — | |
| MRNRe-implementation=false, Extra Captions=false2022.03 | 13.4 | 40 | 61.9 | 38.4 | — | |
| MOON2.0Setting=Zero-shot2026.04 | 13.1 | 27.09 | — | — | 23.16 | |
| MOON2.0Setting=Zero-shot2026.04 | 13.05 | 31.39 | — | — | 25.25 | |
| Relationship2018.12 | 13 | 40.5 | 62.4 | — | — | |
| Relationship2020.06 | 13 | 40.5 | 62.4 | — | — | |
| RelationshipRe-implementation=false, Extra Captions=false2022.03 | 13 | 40.5 | 62.4 | 38.6 | — | |
| FiLM2018.12 | 12.9 | 39.5 | 61.9 | — | — | |
| FiLM2020.06 | 12.9 | 39.5 | 61.9 | — | — | |
| FiLMRe-implementation=false, Extra Captions=false2022.03 | 12.9 | 39.5 | 61.9 | 38.1 | — | |
| Show and Tell2018.12 | 12.3 | 40.2 | 61.8 | — | — | |
| Show and Tell2020.06 | 12.3 | 40.2 | 61.8 | — | — | |
| Param Hashing2018.12 | 12.2 | 40 | 61.7 | — | — | |
| Param Hashing2020.06 | 12.2 | 40 | 61.7 | — | — | |
| Concatenation2018.12 | 11.9 | 39.7 | 62.6 | — | — | |
| MOONSetting=Zero-shot2026.04 | 11.71 | 25.09 | — | — | 20.02 | |
| MOONSetting=Zero-shot2026.04 | 10.82 | 27.73 | — | — | 22.89 | |
| Qwen3-VL-2B (SFT)Setting=SFT (Supervised Fine-Tuning)2026.04 | 10.68 | 26.83 | — | — | 22.79 | |
| GME-Qwen2VLSetting=Zero-shot2026.04 | 10.31 | 27.23 | — | — | 21.14 | |
| GME-Qwen2VLSetting=Zero-shot2026.04 | 8.13 | 26.06 | — | — | 20.31 | |
| Han et al.2018.12 | 6.3 | 19.9 | 38.3 | — | — | |
| Han et al.2020.06 | 6.3 | 19.9 | 38.3 | — | — | |
| MM-EmbedSetting=Zero-shot2026.04 | 6.12 | 22.57 | — | — | 15.91 | |
| MM-EmbedSetting=Zero-shot2026.04 | 5.97 | 22.25 | — | — | 15.71 | |
| Qwen3-VL-2B (SFT)Setting=SFT (Supervised Fine-Tuning)2026.04 | 5.16 | 18.85 | — | — | 13.3 | |
| InternVL3.5-2B (SFT)Setting=SFT (Supervised Fine-Tuning)2026.04 | 5.04 | 17.27 | — | — | 12.39 | |
| FashionCLIPSetting=Zero-shot2026.04 | 4.92 | 15.14 | — | — | 11.67 | |
| CASLIE-SSetting=Zero-shot2026.04 | 4.71 | 14.12 | — | — | 11.25 | |
| CASLIE-SSetting=Zero-shot2026.04 | 4.41 | 13.89 | — | — | 10.04 | |
| SigLIP2Setting=Zero-shot2026.04 | 4.39 | 14.7 | — | — | 10.49 | |
| FashionCLIPSetting=Zero-shot2026.04 | 4.34 | 12.97 | — | — | 9.15 | |
| SigLIP2Setting=Zero-shot2026.04 | 4.19 | 13.9 | — | — | 10.03 | |
| Image only2018.12 | 3.5 | 22.7 | 43.7 | — | — | |
| InternVL3.5-2B (SFT)Setting=SFT (Supervised Fine-Tuning)2026.04 | 3.3 | 11.97 | — | — | 8.37 | |
| Qwen3-VL-EmbeddingSetting=Zero-shot2026.04 | 2.56 | 9.76 | — | — | 6.65 | |
| Text only2018.12 | 1 | 12.3 | 21.8 | — | — | |
| InternVL3.5-2BSetting=Zero-shot2026.04 | 0.58 | 3.14 | — | — | 1.96 | |
| Qwen3-VL-2BSetting=Zero-shot2026.04 | 0.49 | 2.57 | — | — | 1.59 | |
| InternVL3.5-2BSetting=Zero-shot2026.04 | 0.36 | 2.6 | — | — | 1.49 | |
| Qwen3-VL-2BSetting=Zero-shot2026.04 | 0.08 | 0.3 | — | — | 0.19 |