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 Conformer-Transducer XXLarge
Model
NVIDIA
NVIDIA
STT En Conformer-Transducer XXLarge

Conformer-Transducer-XXLarge model for English Automatic Speech Recognition, trained on NeMo ASRSET

1 Version
1.8.0Selected
04/14/2022 11:56 PM UTC3.44 GB
Accuracy
KeyValue
NSC Part 16.64 %
Librispeech test-other3.14 %
Librispeech dev-other3.09 %
WSJ Eval 921.49 %
WSJ Dev 932.42 %
Librispeech dev-clean1.52 %
Multilingual Librispeech dev (EN)5.29 %
Librispeech test-clean1.72 %
Multilingual Librispeech test (EN)5.85 %
Model
KeyValue
Encoder Dimension1024
Number of Encoder Layers38
DatasetNeMo ASRSET 2.0
ARCHITECTUREConformer-Transducer
Number of Predictor Layers2
INPUTS16000 KHZ MONO-CHANNEL AUDIO (WAV FILES)
Number of Weights1B
OUTPUTSTRANSCRIBED SPEECH