Bentham dataset results
| Model | CER | WER | WER BoW |
|---|---|---|---|
| The Text Titan I | 7.07% | 12.41% | 8.54% |
| Gpt-4o-mini-2024-07-18 | 9.48% | 15.09% | 13.16% |
| Gpt-4o-2024-08-06 | 16.62% | 20.73% | 18.89% |
| Claude-3-5-sonnet-20240620 | 10.97% | 14.46% | 12.24% |
| MinicpmV-2 6 | 11.76% | 17.24% | 13.91% |
| Qwen2-VL-7B | 8.01% | 12.94% | 11% |
| Pixtral-12B | 28.08% | 38.25% | 30.32% |
| InternVL2-8B | 76.81% | 95.92% | 81.67% |
| Phi-3-mini-128k-instruct | 32.03% | 41.85% | 38.73% |
| Model | CER | WER | WER BoW |
|---|---|---|---|
| The Text Titan I | |||
| Gpt-4o-mini-2024-07-18 | 9.48% | 15.09% | 13.16% |
| Gpt-4o-2024-08-06 | 16.62% | 20.73% | 18.89% |
| Claude-3-5-sonnet-20240620 | 10.97% | 14.46% | 12.24% |
| MinicpmV-2 6 | 11.76% | 17.24% | 13.91% |
| Qwen2-VL-7B | 8.01% | 12.94% | 11% |
| Pixtral-12B | 28.08% | 38.25% | 30.32% |
| InternVL2-8B | 76.81% | 95.92% | 81.67% |
| Phi-3-mini-128k-instruct | 32.03% | 41.85% | 38.73% |