RIMES dataset results
| Model | CER | WER | WER BoW |
|---|---|---|---|
| The Text Titan I | 10.71% | 20.76% | 17.93% |
| Gpt-4o-mini-2024-07-18 | 3.22% | 7.68% | 6.29% |
| Gpt-4o-2024-08-06 | 1.69% | 3.76% | 3.30% |
| Claude-3-5-sonnet-20240620 | 1.63% | 4% | 3.36% |
| MinicpmV-2 6 | 7.98% | 16.55% | 15% |
| Qwen2-VL-7B | 6.47% | 12.62% | 10.42% |
| Pixtral-12B | 13.02% | 23.55% | 18.73% |
| InternVL2-8B | 35.19% | 66.53% | 57.61% |
| Phi-3-mini-128k-instruct | 23.58% | 48.45% | 43.60% |
| Model | CER | WER | WER BoW |
|---|---|---|---|
| The Text Titan I | 10.71% | 20.76% | 17.93% |
| Gpt-4o-mini-2024-07-18 | 3.22% | 7.68% | 6.29% |
| Gpt-4o-2024-08-06 | 1.69% | ||
| Claude-3-5-sonnet-20240620 | 4% | 3.36% | |
| MinicpmV-2 6 | 7.98% | 16.55% | 15% |
| Qwen2-VL-7B | 6.47% | 12.62% | 10.42% |
| Pixtral-12B | 13.02% | 23.55% | 18.73% |
| InternVL2-8B | 35.19% | 66.53% | 57.61% |
| Phi-3-mini-128k-instruct | 23.58% | 48.45% | 43.60% |