Visual Question Answering on TextVQA (test) using Generation Metrics
0.71CIDErVCD
Evaluation Results
| Method | Links | |||||
|---|---|---|---|---|---|---|
| VCDBackbone=InstructBLIP2024.03 | 0.71 | 0.36 | 0.32 | 0.3 | 0.31 | |
| ICDBackbone=InstructBLIP2024.03 | 0.69 | 0.35 | 0.3 | 0.29 | 0.3 | |
| InstructBLIP2024.03 | 0.56 | 0.29 | 0.22 | 0.19 | 0.19 |