Counting Visual Question Answering on VQA Count-QA
1.8mRMSEens
Evaluation results
| Method | Links | |
|---|---|---|
| ensDescription=Combination of glance-ft-1L, aso-sub-ft-1L-3x3 and seq-sub-ft-3x32016.04 | 1.8 | |
| seq-sub-ft-3x3Mode=Fine-tuned, Grid=3x32016.04 | 1.81 | |
| aso-sub-ft-1L-3x3Mode=Fine-tuned, Hidden Layers=1, Grid=3x32016.04 | 1.94 | |
| glance-ft-1LMode=Fine-tuned, Hidden Layers=12016.04 | 2.19 | |
| Deeper LSTM2016.04 | 2.71 | |
| detect2016.04 | 2.72 | |
| SOTA VQA2016.04 | 3.25 |