Arabic Text Diacritization On Catt Dataset
Metrics
DER(%)
WER (%)
Results
Performance results of various models on this benchmark
Model Name | DER(%) | WER (%) | Paper Title | Repository |
|---|---|---|---|---|
| CBHG | 10.808 | 42.680 | CATT: Character-based Arabic Tashkeel Transformer | |
| Multilevel Diacritizer | 16.482 | 60.844 | CATT: Character-based Arabic Tashkeel Transformer | |
| CATT ED | 8.624 | 34.191 | CATT: Character-based Arabic Tashkeel Transformer | |
| Sakhr | 13.841 | 56.661 | CATT: Character-based Arabic Tashkeel Transformer | |
| Command R+ | 13.169 | 48.518 | CATT: Character-based Arabic Tashkeel Transformer | |
| Shakkala | 13.494 | 50.387 | CATT: Character-based Arabic Tashkeel Transformer | |
| Deep Diacritization (D2) | 13.310 | 49.417 | Deep Diacritization: Efficient Hierarchical Recurrence for Improved Arabic Diacritization | |
| Deep Diacritization (D3) | 58.313 | 98.710 | Deep Diacritization: Efficient Hierarchical Recurrence for Improved Arabic Diacritization | |
| GPT-4 | 9.515 | 38.311 | CATT: Character-based Arabic Tashkeel Transformer | |
| Farasa Diacritization | 17.825 | 65.783 | - | - |
| Alkhalil | 14.232 | 53.413 | CATT: Character-based Arabic Tashkeel Transformer | |
| CATT EO | 8.762 | 35.597 | CATT: Character-based Arabic Tashkeel Transformer |
0 of 12 row(s) selected.