Model
Conformer-Transducer-Large model for Italian Automatic Speech Recognition, Trained on NeMo ASRSET 2.0
Use the NGC CLI to download:
Copied!
1 Version
1.13.0Selected11/01/2022 10:37 PM UTC460.16 MB Copied!
1.13.0Selected
11/01/2022 10:37 PM UTC460.16 MB
Copied!
Accuracy
| Key | Value |
|---|---|
| MLS test | 12.18 |
| MCV 11.0 dev | 4.80 |
| MLS dev | 14.62 |
| VoxPopuli dev | 12.00 |
| VoxPopuli test | 15.15 |
| MCV 11.0 test | 5.24 |
Model
| Key | Value |
|---|---|
| Architecture | CONFORMER TRANSDUCER |
| Outputs | Transcribed speech |
| Inputs | 16 kHz Mono-channel |