CatalogModels
Models
The NGC catalog offers 100s of pre-trained models for computer vision, speech, recommendation, and more. Bring AI faster to market by using these models as-is or quickly build proprietary models with a fraction of your custom data.
Sort: Last Modified
English Citrinet ASR model trained on ASR set 3.0, no-weight-decay
ESS is a DNN that estimates disparity for a stereo image pair and returns a continuous disparity map for the given left image.
Bi3D is a binary depth classification network that is used to classify the depth of objects at a given distance.
English Conformer-XL ASR model trained on ASR set 4.0
TTS En VITS
Model
End-to-end parallel speech synthesis model
wav2vec 2.0 base PyTorch checkpoint pre-trained on LibriSpeech 960 h, fine-tuned on LibriSpeech 960 h
wav2vec 2.0 Base PyTorch checkpoint pre-trained on LibriSpeech 960 h
MoFlow PyTorch checkpoint
QuartzNet15x5 model trained on WSJ, LibriSpeech and Mozilla's Common Voice En with NeMo
This collection contains two models. 1) Multi-speaker 44100Hz FastPitch trained on approximately 20 hours of Latin American Spanish speech from 174 speakers. 2) HiFiGAN trained on mel spectrograms produced by the Multi-speaker FastPitch in (1).
Re-Identification
Model
Re-Identification network to generate embeddings for identifying persons in different scenes.
Spanish Conformer ASR model trained on ASR set 2.0.
For each word in the input text, the model: 1) predicts a punctuation mark that should follow the word (if any), the model supports commas, periods and question marks) and 2) predicts if the word should be capitalized or not.
Speech to Text Jasper model for English.
English Quartznet ASR model trained on ASR set 1.2
English ASR model trained on ASR Set 1.2, Noise Robust
Base Mandarin grammar
HifiGAN is a neural vocoder model for text-to-speech applications. It is intended as the second part of a two-stage speech synthesis pipeline, with a mel-spectrogram generator such as FastPitch as the first stage.
Riva multisepaker with IPA for G2P
Mel-Spectrogram prediction conditioned on input text with LJSpeech voice.
FastPitch is a mel-spectrogram generator, designed to be used as the first part of a neural text-to-speech system in conjunction with a neural vocoder
Riva multisepaker with IPA for G2P
Contains files used in rmir creation
GAN-based waveform generator from mel-spectrograms.
A pre-trained model for volumetric (3D) segmentation of the spleen from CT image.