Fill-In-The-Blank on Polyvore Disjoint (test)
64.17FITB AccuracyMUFIN
Evaluation Results
| Method | Links | |
|---|---|---|
| MUFINα=12023.09 | 64.17 | |
| MUFINBackbone=Resnet18, α=12023.09 | 61.63 | |
| ADDE-OModality=Visual + Textual2023.09 | 60.53 | |
| CSA-NetLabels α%=100%, Uses text labels=false, Explicit type conditioning=true2021.09 | 59.3 | |
| CSA-NetModality=Visual2023.09 | 59.26 | |
| Type-awareModality=Visual + Textual2023.09 | 55.65 | |
| Type-AwareLabels α%=100%, Uses text labels=true, Explicit type conditioning=true2021.09 | 55.2 | |
| OursLabels α%=100%, Uses text labels=false, Explicit type conditioning=false2021.09 | 54.6 | |
| S-VALModality=Visual2023.09 | 54.3 | |
| SCE-Net averageModality=Visual + Textual2023.09 | 53.67 | |
| SCE-Net (avg)Labels α%=100%, Uses text labels=true, Explicit type conditioning=false2021.09 | 53.6 | |
| CSN T1:1 + VSELabels α%=100%, Uses text labels=true, Explicit type conditioning=true2021.09 | 53 | |
| CSN T1:1Labels α%=100%, Uses text labels=false, Explicit type conditioning=true2021.09 | 52.5 | |
| Siamese NetworkLabels α%=100%, Uses text labels=false, Explicit type conditioning=false2021.09 | 51.8 | |
| Ours + DLabels α%=5%, Uses text labels=false, Explicit type conditioning=false, Additional unlabeled data=true2021.09 | 51.5 | |
| SSVRModality=Visual + Textual2023.09 | 51.5 | |
| OursLabels α%=5%, Uses text labels=false, Explicit type conditioning=false2021.09 | 51.4 | |
| Siamese NetworkLabels α%=5%, Uses text labels=false, Explicit type conditioning=false2021.09 | 47 | |
| Bi-LSTM +VSELabels α%=100%, Uses text labels=true, Explicit type conditioning=false2021.09 | 39.4 | |
| Color attributeLabels α%=N/A, Uses text labels=false, Explicit type conditioning=false2021.09 | 39.2 |