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: Most Popular
Displaying 0 models
NVIDIA AI Enterprise Support
91
Use Case
61
58
18
15
13
13
5
5
1
NVIDIA Platform
103
96
82
41
25
24
1
Framework
Industry
Solution
Publisher
Language
Other
Japanese (ja-JP) Conformer ASR model trained on ASR set 4.0
Italian (it-IT) Conformer ASR model trained on ASR set 1.0
For each word in the input text, the model: predicts a punctuation mark that should follow the word (if any).
For each word in the input text, the model: predicts a punctuation mark that should follow the word (if any).
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.
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.
EfficientDet-D0 Pytorch checkpoint trained on COCO using batchsize=1200
Intent and Slot classification of the queries for the misty bot with DistilBert model trained on weather, smalltalk and POI (places of interest) data.
Base Mandarin grammar
Hindi Citrinet ASR model trained on RIVA ASR set
Brazilian Portuguese (pt-BR) Citrinet-1024 ASR model trained on ASR set 1.0
Spanish EMEA (es-ES) Citrinet-1024 ASR model trained on ASR set 1.0
Japanese (ja-JP) Conformer ASR model trained on ASR set 3.0
Neural VAD model used in Riva Speaker Diarization
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.
The model identifies a category/entity the word in the input text belongs to.
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, hyphens and question marks) and 2) predicts if the word should be capitalized or not.
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.
For each word in the input text, the model predicts a punctuation mark that should follow the word (if any), the model supports commas, poornvirams, exclaimation marks and question marks.
Wide&Deep Base TensorFlow2 checkpoint trained with AMP on NVTabular preprocessed dataset with multihot embeddings
BART PyT checkpoint for summarization on XSum dataset
BERT Large PaddlePaddle checkpoint pretrained with LAMB optimizer using AMP
BERT Distilled 4L-288D PyTorch checkpoint distilled on SQuAD v1.1 dataset using AMP
BERT Distilled 6L-768D PyTorch Phase2 checkpoint pretrained using 67K steps on seqLen128 and 6k steps on seqLen512