Visual Question Answering on InfographicVQA (test)
80.3ANLSGemini Ultra 1.0
Evaluation results
| Method | Links | |
|---|---|---|
| Gemini Ultra 1.0open-source=false2024.04 | 80.3 | |
| Gemini Pro 1.0open-source=false2024.04 | 75.2 | |
| Qwen-VL-Maxopen-source=false2024.04 | 73.4 | |
| Gemini Pro 1.5open-source=false2024.04 | 72.7 | |
| InternVL 1.5#param=26B, open-source=false2024.04 | 72.5 | |
| LLaVA-NeXT#param=35B, open-source=false2024.04 | 51.5 | |
| DocOwl-1.5#param=8B, open-source=true2024.04 | 50.7 | |
| UDOPtype=State of the art w/ pipelines2022.10 | 47.4 | |
| Pix2Struct LargePretraining=Screenshot parsing, Pixel only=true2022.10 | 40 | |
| InternVL 1.2#param=40B, open-source=false2024.04 | 39.5 | |
| Pix2Struct BasePretraining=Screenshot parsing, Pixel only=true2022.10 | 38.2 | |
| Text-Monkey#param=10B, open-source=true2024.04 | 28.6 | |
| DonutPretraining=OCR, Pixel only=true2022.10 | 11.6 |