Visual Question Answering Vqa On Infoseek
評価指標
Accuracy
評価結果
このベンチマークにおける各モデルのパフォーマンス結果
モデル名 | Accuracy | Paper Title | Repository |
|---|---|---|---|
| PaLI-X | 24 | PaLI-X: On Scaling up a Multilingual Vision and Language Model | |
| BLIP2 | 14.6 | BLIP-2: Bootstrapping Language-Image Pre-training with Frozen Image Encoders and Large Language Models | |
| InstructBLIP | 14.5 | - | - |
| CLIP + FiD | 20.9 | Can Pre-trained Vision and Language Models Answer Visual Information-Seeking Questions? | - |
| CLIP + PaLM (540B) | 20.4 | Can Pre-trained Vision and Language Models Answer Visual Information-Seeking Questions? | - |
| PaLI | 19.7 | Can Pre-trained Vision and Language Models Answer Visual Information-Seeking Questions? | - |
| RA-VQAv2 w/ PreFLMR | 30.65 | PreFLMR: Scaling Up Fine-Grained Late-Interaction Multi-modal Retrievers | - |
0 of 7 row(s) selected.