ContainersNVIDIADeepSeek-R1NVIDIALlama-3.1-Nemotron-70B-InstructNVIDIAPyTorch
CollectionsNVIDIAOmniverse Kit (FB)NVIDIADeepStream SDKNVIDIAOmniverse Kit App Streaming
ModelsNVIDIAStyleGAN3 pretrained modelsNVIDIAPeopleNetNVIDIATrafficCamNet
ResourcesNVIDIARiva Skills Quick Start
Helm ChartsNVIDIAGPU OperatorNVIDIANVIDIA NIM Operator
NVIDIA
NVIDIA
STT En ContextNet 512
Model
NVIDIA
NVIDIA
STT En ContextNet 512

ContextNet-512 model for English Automatic Speech Recognition, trained on NeMo ASRSET.

3 Versions
1.6.0
01/13/2022 8:00 PM UTC142.63 MB
Accuracy
KeyValue
LibrispeechDevClean2.0 %
LibrispeechDevOther4.8 %
LibrispeechTestClean2.2 %
LibrispeechTestOther5.0 %
MLSDev6.6 %
MLSTest7.3 %
NSCPartOne5.9 %
WSJDev932.8 %
WSJEval921.4 %
Model
KeyValue
ARCHITECTUREContextNet
INPUTS16000 KHZ MONO-CHANNEL AUDIO (WAV FILES)
OUTPUTSTRANSCRIBED SPEECH
1.4.0Selected
09/21/2021 10:20 PM UTC142.25 MB
Accuracy
KeyValue
LibrispeechDevClean2.2 %
LibrispeechDevOther5.3 %
LibrispeechTestClean2.4 %
LibrispeechTestOther5.3 %
NSCPartOne5.7 %
WSJDev933.1 %
WSJEval921.6 %
Model
KeyValue
ARCHITECTUREContextNet
INPUTS16000 KHZ MONO-CHANNEL AUDIO (WAV FILES)
OUTPUTSTRANSCRIBED SPEECH
1.0.0
09/16/2021 12:19 AM UTC141.97 MB
Accuracy
KeyValue
LibrispeechDevClean3.7
LibrispeechDevOther7.9
LibrispeechTestClean3.9
LibrispeechTestOther7.8
NSCPartOne8.5
WSJDev934.2
WSJEval922.8
Model
KeyValue
ARCHITECTUREContextNet
INPUTS16000 KHZ MONO-CHANNEL AUDIO (WAV FILES)
OUTPUTSTRANSCRIBED SPEECH