FLEURS
Benchmarks
| Task Name | Dataset Name | SOTA Result | Trend | |
|---|---|---|---|---|
| Automatic Speech Recognition | Fleurs | WER0.1058 | 56 | |
| Automatic Speech Recognition | FLEURS en | WER2.72 | 49 | |
| Automatic Speech Recognition | Fleurs zh | WER0.83 | 34 | |
| Speech Recognition | Fleurs En | WER2.4 | 24 | |
| Script confusion mitigation | FLEURS zh-trad (test) | Accuracy91 | 21 | |
| Script confusion mitigation | FLEURS zh-sim (test) | Normalized Edit Similarity Accuracy93 | 21 | |
| Script confusion mitigation | FLEURS sr-cyrl (test) | Normalized Edit Similarity0.94 | 21 | |
| Script confusion mitigation | FLEURS sr-latn (test) | Accuracy (Normalized Edit Similarity)96 | 21 | |
| Hallucination detection | FLEURS S2TT (test) | Hallucination Rate (%)89.66 | 18 | |
| Speech Recognition | FLEURS Kazakh (test) | WER (Oracle)15.4 | 17 | |
| Automatic Speech Recognition | fleurs Tamil | WER28.68 | 17 | |
| Automatic Speech Recognition | Fleurs en (test) | WER0.336 | 17 | |
| Speech Translation | FLEURS X→En (test) | BLEU (Average)37.4 | 16 | |
| Noise Robustness | FLEURS (test) | Robustness Score (Gaussian Noise)84.46 | 15 | |
| Speech-to-Text Translation | FLEURS (test) | Avg. Score86.3 | 15 | |
| Speech Translation | FLEURS (test) | COMET (Low)0.853 | 15 | |
| Automatic Speech Recognition | FLEURS | WER (en)2.71 | 13 | |
| Automatic Speech Recognition | FLEURS (test) | Average Error Rate0.32 | 13 | |
| Automatic Speech Recognition | FLEURS | AMH12.77 | 13 | |
| Document Retrieval | Fleurs Retrieval-Eval (test) | nDCG@557.89 | 13 | |
| Speech-to-Text Translation | FLEURS 9×27 directions (test) | ara78.7 | 10 | |
| Automatic Speech Recognition | FLEURS-RO General | WER17.57 | 9 | |
| Automatic Speech Recognition | FLEURS Somali (test) | SFR100 | 9 | |
| Automatic Speech Recognition | FLEURS Malayalam (test) | SFR100 | 9 | |
| Automatic Speech Recognition | FLEURS Bengali (test) | SFR100 | 9 |
Showing 25 of 112 rows