Model
ContextNet-1024 model for English Automatic Speech Recognition, trained on English subset of Multilingual Librispeech Dataset.
Use the NGC CLI to download:
Copied!
1 Version
1.0.0Selected09/16/2021 12:17 AM UTC453.91 MB Copied!
1.0.0Selected
09/16/2021 12:17 AM UTC453.91 MB
Copied!
Accuracy
| Key | Value |
|---|---|
| Librispeech Dev-Other | 4.1 % |
| Librispeech Test-Other | 4.2 % |
| MLS Dev | 4.6 % |
| MLS Test | 5.6 % |
Model
| Key | Value |
|---|---|
| ARCHITECTURE | ContextNet |
| INPUTS | 16000 KHZ MONO-CHANNEL AUDIO (WAV FILES) |
| OUTPUTS | TRANSCRIBED SPEECH |