NLP Overview

Riva supports basic natural language processing (NLP) text and token classification functionality; with classifiers that can be modified and deployed using NVIDIA NeMo. For more information about the Riva NLP API, refer to the riva/proto/riva_nlp.proto section.

Punctuation and captilization

The punctuation and captilization models deployed on Riva expect unpunctuated inputs. Inputs which have punctuation may not produce expected results.

Checking deployed models

Once a server is running retrieving the available models can be done via the GetRivaNLPConfig rpc. For each model availble to make inference requets, the rpc returns the parameters used when the model was deployed.