Phone Alignment on CSJ (test)
78.44Mean Alignment AccuracyBFA*
Evaluation Results
| Method | Links | ||||
|---|---|---|---|---|---|
| BFA*type=Baseline aligner, model_nature=neural aligner2026.06 | 78.44 | 2.19 | 4.9 | 24.69 | |
| Juliustype=Baseline aligner2026.06 | 19.25 | 39.02 | 70.64 | 95.35 | |
| SPPAStype=Baseline aligner2026.06 | 17.75 | 44.43 | 75.17 | 95.53 | |
| MFA English 3.0model_category=MFA 3.0 pretrained, adaptation=none2026.06 | 14.3 | 64.64 | 83.89 | 95.85 | |
| MAUStype=Baseline aligner2026.06 | 13.46 | 62.59 | 84 | 96.4 | |
| MFA English 3.0 +adaptedmodel_category=MFA 3.0 pretrained, adaptation=adapted2026.06 | 11.67 | 66.82 | 86.84 | 97.69 | |
| MFA Japanese 3.0model_category=MFA 3.0 pretrained, adaptation=none2026.06 | 10.82 | 63.44 | 88.34 | 98.81 | |
| MFA Japanese 3.0 +adaptedmodel_category=MFA 3.0 pretrained, adaptation=adapted2026.06 | 10.7 | 63.35 | 88.55 | 98.9 | |
| MFA trained +rulesmodel_category=MFA 3.0 trained on dataset, phonological_rules=added2026.06 | 10.26 | 67.34 | 88.44 | 98.77 | |
| MFA trained -PPmodel_category=MFA 3.0 trained on dataset, pronunciation_probability_estimation=removed2026.06 | 10.18 | 67.56 | 88.54 | 98.81 | |
| MFA trainedmodel_category=MFA 3.0 trained on dataset, pronunciation_probability_estimation=default2026.06 | 10.13 | 67.75 | 88.49 | 98.8 |