Image Captioning on MS-COCO 2014 (test)
78.7BLEU-4OFA
Evaluation results
| Method | Links | |||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| OFAReference count=c402022.08 | 78.7 | 98.1 | 94.4 | 87.8 | 42.7 | 79 | 149.6 | — | — | — | — | |
| GITReference count=c402022.08 | 78.3 | 98.1 | 94.4 | 87.6 | 42.1 | 78.4 | 149.8 | — | — | — | — | |
| ExpansionNet v2Reference count=c402022.08 | 75.3 | 96.9 | 92.6 | 85 | 40.1 | 76.4 | 140.8 | — | — | — | — | |
| PureTReference count=c402022.08 | 74.1 | 96.5 | 91.8 | 83.9 | 39.9 | 75.9 | 138.3 | — | — | — | — | |
| DLCTReference count=c402022.08 | 74 | 96.6 | 91.7 | 83.8 | 39.6 | 75.3 | 135.4 | — | — | — | — | |
| RSTNetReference count=c402022.08 | 73.1 | 96.4 | 91.3 | 83 | 39.1 | 74.6 | 134 | — | — | — | — | |
| GETReference count=c402022.08 | 72.9 | 96.1 | 90.9 | 82.8 | 38.8 | 74.4 | 132.5 | — | — | — | — | |
| X-TransformerReference count=c402022.08 | 72.4 | 95.7 | 90.5 | 82.5 | 39.2 | 75 | 133.5 | — | — | — | — | |
| AoANetReference count=c402022.08 | 71.2 | 95 | 89.6 | 81.3 | 38.5 | 74.5 | 129.6 | — | — | — | — | |
| GCN-LSTMReference count=c402022.08 | 69.7 | — | 89.3 | 80.3 | 37.6 | 73.4 | 126.5 | — | — | — | — | |
| SGAEReference count=c402022.08 | 69.7 | 95.3 | 89.5 | 80.4 | 37.2 | 73.6 | 126.5 | — | — | — | — | |
| Up-DownReference count=c402022.08 | 68.5 | 95.2 | 88.8 | 79.4 | 36.7 | 72.4 | 120.5 | — | — | — | — | |
| OFAReference count=c52022.08 | 43.6 | 84.5 | 70.1 | 55.9 | 32.1 | 62.5 | 147.2 | — | — | — | — | |
| GITReference count=c52022.08 | 43.2 | 84.3 | 70 | 55.7 | 31.9 | 62 | 146.4 | — | — | — | — | |
| ExpansionNet v2Reference count=c52022.08 | 42.1 | 83.3 | 68.8 | 54.4 | 30.4 | 60.8 | 138.5 | — | — | — | — | |
| PureTReference count=c52022.08 | 41.4 | 82.8 | 68.1 | 53.6 | 30.1 | 60.4 | 136 | — | — | — | — | |
| DLCTReference count=c52022.08 | 40.6 | 82.4 | 67.4 | 52.8 | 29.8 | 59.8 | 133.3 | — | — | — | — | |
| X-TransformerReference count=c52022.08 | 40.3 | 81.9 | 66.9 | 52.4 | 29.6 | 59.5 | 131.1 | — | — | — | — | |
| RSTNetReference count=c52022.08 | 40 | 82.1 | 67 | 52.2 | 29.6 | 59.5 | 131.9 | — | — | — | — | |
| GETReference count=c52022.08 | 39.7 | 81.6 | 66.5 | 51.9 | 29.4 | 59.1 | 130.3 | — | — | — | — | |
| AoANetReference count=c52022.08 | 39.4 | 81 | 65.8 | 51.4 | 29.1 | 58.9 | 126.9 | — | — | — | — | |
| GCN-LSTMReference count=c52022.08 | 38.7 | — | 65.5 | 50.8 | 28.5 | 58.5 | 125.3 | — | — | — | — | |
| SGAEReference count=c52022.08 | 38.5 | 81 | 65.6 | 50.7 | 28.2 | 58.6 | 123.8 | — | — | — | — | |
| Ca-CaptionerType=Diffusion2025.12 | 38.3 | — | — | — | 28.8 | 57.7 | 122 | 21.8 | — | — | — | |
| AoAType=Autoregressive2025.12 | 38.1 | 78.7 | — | — | 28.5 | 58.2 | 122.7 | 21.7 | — | — | — | |
| SSTType=Mixed2025.12 | 37.6 | 78.3 | 62.6 | 48.8 | 29.2 | 58.3 | 125.1 | 22.8 | — | — | — | |
| CIICGType=Autoregressive2025.12 | 37.3 | 77.5 | — | — | 28.5 | 57.4 | 119 | 21.5 | — | — | — | |
| SCD-NetType=Diffusion2025.12 | 37.3 | 79 | 63.4 | 49.1 | 28.1 | 58 | 118 | 21.6 | — | — | — | |
| X-TransformerType=Autoregressive2025.12 | 37 | 77.3 | 61.5 | 47.8 | 28.7 | 57.5 | 120 | 21.8 | — | — | — | |
| Up-DownReference count=c52022.08 | 36.9 | 80.2 | 64.1 | 49.1 | 27.6 | 57.1 | 117.9 | — | — | — | — | |
| EENAICType=Diffusion2025.12 | 36.9 | 79.7 | — | — | 27.9 | 58 | 122.6 | — | — | — | — | |
| PureTType=Autoregressive2025.12 | 36.8 | 77.7 | 61.6 | 47.8 | 28.8 | 57.7 | 122.4 | 21.7 | — | — | — | |
| Show, Recall and TellType=Autoregressive2025.12 | 36.6 | 77.1 | — | — | 28 | 56.9 | 116.9 | 21.3 | — | — | — | |
| ExpansionNet v2Type=Autoregressive2025.12 | 36.6 | 77.1 | 61.2 | 47.6 | 29.1 | 57.8 | 123.6 | 22.7 | — | — | — | |
| BUTDType=Autoregressive2025.12 | 36.2 | 77.2 | — | — | 27 | 56.4 | 113.5 | 20.3 | — | — | — | |
| ObjTransType=Autoregressive2025.12 | 35.5 | 76.6 | — | — | 28 | 56.6 | 115.4 | 21.2 | — | — | — | |
| DECapType=Diffusion2025.12 | 35.3 | 78.5 | 62.2 | 47.4 | 29 | 58.4 | 121.2 | 22.7 | — | — | — | |
| DDCapType=Diffusion2025.12 | 34.7 | — | — | — | — | 58 | 116.7 | 21.5 | — | — | — | |
| BitDiffusionType=Diffusion2025.12 | 34.7 | — | — | — | — | 58 | 115 | — | — | — | — | |
| TransformerType=Autoregressive2025.12 | 34 | 76.1 | 59.9 | 45.2 | 27.6 | 56.2 | 113.3 | 21 | — | — | — | |
| Show, Observe and TellType=Autoregressive2025.12 | 33.8 | 74.3 | 57.9 | 44.3 | — | 54.9 | 104.4 | — | — | — | — | |
| Show, Attend and TellType=Autoregressive2025.12 | 25 | 71.8 | 50.4 | 35.7 | 23 | — | — | — | — | — | — | |
| Show and TellType=Autoregressive2025.12 | 24.3 | — | — | — | 23.7 | — | 85.5 | — | — | — | — | |
| Beam SearchBackbone LVLM=LLaVA-1.52026.02 | — | — | — | — | — | — | — | — | 19.5 | 6.84 | 15.99 | |
| Beam SearchBackbone LVLM=MiniGPT-42026.02 | — | — | — | — | — | — | — | — | 30.1 | 11.87 | 15.35 | |
| Beam SearchBackbone LVLM=mPLUG-Owl22026.02 | — | — | — | — | — | — | — | — | 20.3 | 6.92 | 15.43 | |
| DoLaBackbone LVLM=LLaVA-1.52026.02 | — | — | — | — | — | — | — | — | 20.2 | 6.55 | 15.58 | |
| DoLaBackbone LVLM=MiniGPT-42026.02 | — | — | — | — | — | — | — | — | 31 | 12.15 | 14.54 | |
| DoLaBackbone LVLM=mPLUG-Owl22026.02 | — | — | — | — | — | — | — | — | 22.4 | 7.36 | 15.13 | |
| GreedyBackbone LVLM=LLaVA-1.52026.02 | — | — | — | — | — | — | — | — | 20.4 | 7.08 | 15.72 | |
| GreedyBackbone LVLM=MiniGPT-42026.02 | — | — | — | — | — | — | — | — | 32.4 | 12.2 | 14.57 | |
| GreedyBackbone LVLM=mPLUG-Owl22026.02 | — | — | — | — | — | — | — | — | 22.9 | 8.62 | 15.01 | |
| HALCBackbone LVLM=LLaVA-1.52026.02 | — | — | — | — | — | — | — | — | 16.9 | 5.72 | 16.02 | |
| HALCBackbone LVLM=MiniGPT-42026.02 | — | — | — | — | — | — | — | — | 25.2 | 9.42 | 14.91 | |
| HALCBackbone LVLM=mPLUG-Owl22026.02 | — | — | — | — | — | — | — | — | 18.8 | 7 | 15.33 | |
| HIMEBackbone LVLM=LLaVA-1.52026.02 | — | — | — | — | — | — | — | — | 13.8 | 4.56 | 15.82 | |
| HIMEBackbone LVLM=MiniGPT-42026.02 | — | — | — | — | — | — | — | — | 16.8 | 7.43 | 14.62 | |
| HIMEBackbone LVLM=mPLUG-Owl22026.02 | — | — | — | — | — | — | — | — | 15.4 | 5.19 | 15.81 | |
| HIMEBackbone LVLM=QWen2-VL-8B-Instruct2026.02 | — | — | — | — | — | — | — | — | 17.2 | 4.43 | 11.3 | |
| HIMEBackbone LVLM=QWen3-VL-8B-Instruct2026.02 | — | — | — | — | — | — | — | — | 6 | 3.44 | 8.89 | |
| LUREBackbone LVLM=LLaVA-1.52026.02 | — | — | — | — | — | — | — | — | 19.48 | 6.5 | 15.97 | |
| LUREBackbone LVLM=MiniGPT-42026.02 | — | — | — | — | — | — | — | — | 27.88 | 10.2 | 15.03 | |
| LUREBackbone LVLM=mPLUG-Owl22026.02 | — | — | — | — | — | — | — | — | 21.27 | 7.67 | 15.65 | |
| NulluBackbone LVLM=LLaVA-1.52026.02 | — | — | — | — | — | — | — | — | 15.2 | 5.3 | 15.69 | |
| NulluBackbone LVLM=MiniGPT-42026.02 | — | — | — | — | — | — | — | — | 21.4 | 8.99 | 14.81 | |
| NulluBackbone LVLM=mPLUG-Owl22026.02 | — | — | — | — | — | — | — | — | 15.6 | 5.77 | 15.45 | |
| OPERABackbone LVLM=LLaVA-1.52026.02 | — | — | — | — | — | — | — | — | 17.5 | 6.07 | 16.02 | |
| OPERABackbone LVLM=MiniGPT-42026.02 | — | — | — | — | — | — | — | — | 29.7 | 11.9 | 14.82 | |
| OPERABackbone LVLM=mPLUG-Owl22026.02 | — | — | — | — | — | — | — | — | 20.67 | 6.4 | 15.41 | |
| QWen2-VL-8B-Instruct(QWen2)Mode=Baseline2026.02 | — | — | — | — | — | — | — | — | 20.8 | 5.36 | 11.16 | |
| QWen3-VL-8B-Instruct(QWen3)Mode=Baseline2026.02 | — | — | — | — | — | — | — | — | 8.4 | 4.62 | 9.81 | |
| VCDBackbone LVLM=LLaVA-1.52026.02 | — | — | — | — | — | — | — | — | 20.3 | 7.28 | 14.53 | |
| VCDBackbone LVLM=MiniGPT-42026.02 | — | — | — | — | — | — | — | — | 29.2 | 11.69 | 14.42 | |
| VCDBackbone LVLM=mPLUG-Owl22026.02 | — | — | — | — | — | — | — | — | 22.8 | 8.67 | 15.14 | |
| WoodpeckerBackbone LVLM=LLaVA-1.52026.02 | — | — | — | — | — | — | — | — | 23.85 | 7.5 | 15.01 | |
| WoodpeckerBackbone LVLM=MiniGPT-42026.02 | — | — | — | — | — | — | — | — | 30.4 | 12 | 15 | |
| WoodpeckerBackbone LVLM=mPLUG-Owl22026.02 | — | — | — | — | — | — | — | — | 26.33 | 9.43 | 15.03 |