Handwritten Text Recognition on LAM 34 (test)
2.8CERHTR-VT
Evaluation results
| Method | Links | ||
|---|---|---|---|
| HTR-VTParam.=53.5M2024.09 | 2.8 | 7.4 | |
| OrigamiNet-24Param.=115.3M, Re-implementation=LAM [34]2024.09 | 3 | 11 | |
| OrigamiNet-12Param.=39.0M, Re-implementation=LAM [34]2024.09 | 3.1 | 11.2 | |
| OrigamiNet-18Param.=77.1M, Re-implementation=LAM [34]2024.09 | 3.1 | 11.1 | |
| TrOCRParam.=385.0M, Extra training data=true, Re-implementation=LAM [34]2024.09 | 3.6 | 11.6 | |
| CRNNParam.=18.2M, Re-implementation=LAM [34]2024.09 | 3.8 | 12.9 | |
| GFCNParam.=1.4M, Re-implementation=LAM [34]2024.09 | 5.2 | 18.5 | |
| ViT + DropKeyParam.=37M2024.09 | 5.7 | 16.5 | |
| CNN + BLSTMParam.=9.3M, Re-implementation=LAM [34]2024.09 | 5.8 | 18.4 | |
| DeiTParam.=6M2024.09 | 5.9 | 18.7 | |
| ViTParam.=37M2024.09 | 6.1 | 19.1 | |
| TransformersParam.=54.7M, Re-implementation=LAM [34]2024.09 | 10.2 | 22 |