Integrating Conversational AI Models into Riva

While the TAO Toolkit is an excellent resource to train and finetune models, Riva provides resources to deploy those models into scalable services running on GPUs.

Along with the TAO conversational AI package, we provide the following sample resources on NGC to capture the end to end workflow of training a model with TAO and deploying them to Riva.

Conversational AI Task

Jupyter Notebooks

Speech to Text

Speech to Text Notebook

Speech to Text Citrinet

Speech to Text Citrinet Notebook

Question Answering

Question Answering Notebook

Text Classification

Text Classification Notebook

Token Classification

Token Classification Notebook

Punctuation and Capitalization

Punctuation Capitalization Notebook

Intent and Slot Classification

Intent Slot Classification Notebook

NGram LM Notebook

NGram LM Notebook

Text to Speech

Text to Speech Notebook

Each sample resouce contain 2 sample notebooks,

  1. To train the respective model using TAO Toolkit and generate an exported .riva

  2. To use this exported .riva file and deploy it to Riva.

You may find more information about the same in the Riva Documentation.