Visual Question Answering on CLEVR-CoGenT (val)
99.8AccuracyNS-VQA
Evaluation Results
| Method | Links | ||||
|---|---|---|---|---|---|
| NS-VQAExtra supervision=Programs2024.11 | 99.8 | — | — | — | |
| MDETRExtra supervision=Bound. Box2024.11 | 99.8 | — | — | — | |
| IPRMExtra supervision=None2024.11 | 99.1 | — | — | — | |
| MACExtra supervision=None2024.11 | 99 | — | — | — | |
| FiLMExtra supervision=None2024.11 | 98.3 | — | — | — | |
| PG+EEExtra supervision=Programs2024.11 | 96.6 | — | — | — | |
| IPRMExtra supervision=None2024.11 | 80.3 | — | — | — | |
| FiLMExtra supervision=None2024.11 | 78.8 | — | — | — | |
| MACExtra supervision=None2024.11 | 78.3 | — | — | — | |
| MDETRExtra supervision=Bound. Box2024.11 | 76.7 | — | — | — | |
| PG+EEExtra supervision=Programs2024.11 | 73.7 | — | — | — | |
| NS-VQAExtra supervision=Programs2024.11 | 63.9 | — | — | — | |
| FiLMPre-training=Visual Pre-training2022.01 | — | — | — | 78.8 | |
| LexLSTMPre-training=Seq2Seq Baseline2022.01 | — | — | — | 62.1 | |
| NS-VQAProgram=supervised2018.12 | — | 99.8 | 63.9 | — | |
| NSVQAPre-training=Visual Pre-training2022.01 | — | — | — | 63.9 | |
| S-MACPre-training=Visual Pre-training2022.01 | — | — | — | 78.7 | |
| T5Pre-training=Seq2Seq Baseline2022.01 | — | — | — | 79.7 | |
| VQATransformerPre-training=None2022.01 | — | — | — | 73.3 | |
| VQATransformer + LexSymPre-training=None2022.01 | — | — | — | 85.9 | |
| VQATransformer + SubstitutePre-training=None2022.01 | — | — | — | 84.4 | |
| XNM-DetProgram=supervised2018.12 | — | 98.2 | 72.1 | — | |
| XNM-DetProgram=GT2018.12 | — | 98.3 | 72.2 | — | |
| XNM-GTProgram=supervised2018.12 | — | 99.9 | 99.9 | — | |
| XNM-GTProgram=GT2018.12 | — | 100 | 100 | — |