NVIDIA
NVIDIA
STT En ContextNet 1024 MLS
Model
NVIDIA
NVIDIA
STT En ContextNet 1024 MLS

ContextNet-1024 model for English Automatic Speech Recognition, trained on English subset of Multilingual Librispeech Dataset.

1 Version
1.0.0Selected
09/16/2021 12:17 AM UTC453.91 MB
Accuracy
KeyValue
Librispeech Dev-Other4.1 %
Librispeech Test-Other4.2 %
MLS Dev4.6 %
MLS Test5.6 %
Model
KeyValue
ARCHITECTUREContextNet
INPUTS16000 KHZ MONO-CHANNEL AUDIO (WAV FILES)
OUTPUTSTRANSCRIBED SPEECH