ASR Accent Adaptation on IndicTTS Database Gujarati
9.4WERDITTO (FLMI)
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| DITTO (FLMI)subset_selection_method=FLMI, budget=1230s2021.10 | 9.4 | 100 | |
| DITTO (GCMI)subset_selection_method=GCMI, budget=1230s2021.10 | 9.7 | 100 | |
| Log Determinantsubset_selection_method=LogDet, budget=1230s2021.10 | 10.7 | 24.6 | |
| Facility Locationsubset_selection_method=FL, budget=1230s2021.10 | 10.9 | 22.6 | |
| Random Selectionsubset_selection_method=Random, budget=1230s2021.10 | 11 | 15.6 | |
| Phone Diversitysubset_selection_method=PhoneDiv, budget=1230s2021.10 | 11.1 | 1.5 | |
| Maximum ASR Entropysubset_selection_method=Entropy, budget=1230s2021.10 | 11.2 | 28.8 | |
| Pre-trained Baselinesubset_selection_method=None2021.10 | 13.7 | — |