Visual Dialog On Visual Dialog V1 0 Test Std
评估指标
MRR (x 100)
Mean
NDCG (x 100)
R@1
R@10
R@5
评测结果
各个模型在此基准测试上的表现结果
| Paper Title | Repository | |||||||
|---|---|---|---|---|---|---|---|---|
| MRR ensemble (Naive) | 71.24 | 2.96 | 64.04 | 58.27 | 94.45 | 87.55 | - | - |
| Ensemble FGA + BERT | 70.95 | 2.91 | 67.09 | 57.07 | 95.08 | 88.42 | - | - |
| Two-Step(refactor) | 70.41 | 3.66 | 72.16 | 58.17 | 90.83 | 83.85 | - | - |
| 2 Step: Factor Graph Attention + VD-Bert | 69.92 | 3.84 | 72.83 | 58.3 | 89.6 | 81.55 | Ensemble of MRR and NDCG models for Visual Dialog | |
| 5xFGA (F-RCNNx101) | 69.3 | 3.14 | 57.20 | 55.65 | 94.05 | 86.73 | Factor Graph Attention | |
| CAF | 68.16 | 3.3 | 63.94 | 54.67 | 93.1 | 84.95 | - | - |
| test1 | 67.5 | 3.32 | 63.87 | 53.85 | 93.25 | 84.67 | - | - |
| w/ VQA + CC, single model | 67.5 | 3.32 | 63.87 | 53.85 | 93.25 | 84.67 | - | - |
| sh101 | 67.49 | 3.31 | 63.75 | 53.75 | 93.25 | 85.02 | - | - |
| SCL_48 | 66.63 | 3.41 | 60.91 | 52.52 | 92.27 | 84.1 | - | - |
| Transformer+2cons | 66.53 | 3.4 | 60.33 | 52.62 | 92.5 | 84.12 | - | - |
| single model | 66.2 | 3.25 | 59.33 | 51.62 | 93.7 | 85.05 | - | - |
| Bert2constraints | 65.7 | 3.68 | 58.51 | 51.73 | 91.97 | 82.97 | - | - |
| single-model | 64.95 | 3.44 | 60.31 | 50.48 | 93.15 | 83.15 | - | - |
| MVAN | 64.84 | 3.97 | 59.37 | 51.45 | 90.65 | 81.12 | Multi-View Attention Network for Visual Dialog | |
| jiuyigedian | 64.79 | 3.98 | 58.25 | 51.32 | 90.38 | 81.0 | - | - |
| CARE(Single Model) | 64.62 | 4.29 | 64.79 | 51.82 | 89.95 | 80.35 | - | - |
| gr | 64.58 | 4.03 | 59.23 | 51.25 | 90.05 | 80.92 | - | - |
| clean_wac_4freeze | 64.57 | 3.67 | 57.6 | 49.75 | 91.67 | 82.23 | - | - |
| disc | 64.43 | 4.13 | 58.19 | 50.7 | 90.18 | 80.83 | - | - |
0 of 80 row(s) selected.