Visual Question Answering on GQA 2019 (test)
91.2Binary AccuracyHuman
Evaluation results
| Method | Links | |||||||
|---|---|---|---|---|---|---|---|---|
| Human2021.10 | 91.2 | 87.4 | 98.4 | 97.2 | 98.9 | — | 89.3 | |
| ProToSignal=Prog2021.10 | 79.12 | 51.45 | 93.45 | 86.12 | 96.52 | 3.66 | 65.14 | |
| NSMSignal=SG2021.10 | 78.94 | 49.25 | 93.25 | 84.28 | 96.41 | 3.71 | 63.17 | |
| MMNSignal=Prog2021.10 | 78.9 | 44.89 | 92.49 | 84.55 | 96.19 | 5.54 | 60.83 | |
| PVRSignal=Prog2021.10 | 78.02 | 43.75 | 91.43 | 84.77 | 96.5 | 6 | 59.81 | |
| LXMERTSignal=QA2021.10 | 77.16 | 45.47 | 89.59 | 84.53 | 96.35 | 5.69 | 60.33 | |
| SNMNSignal=Prog2021.10 | 73.4 | 40.82 | 85.11 | 84.79 | 96.37 | 5.14 | 56.09 | |
| MACSignal=QA2021.10 | 71.23 | 38.91 | 81.59 | 84.48 | 96.16 | 5.34 | 54.06 | |
| BottomUpSignal=QA2021.10 | 66.64 | 34.83 | 78.71 | 84.57 | 96.18 | 5.98 | 49.74 |