Toggle in-page Table of Contents
Quick Start Guide
This is the starting point to try out Riva. Specifically, this Quick Start Guide enables you to deploy pretrained models on a local workstation and run a sample client.
Riva Speech AI Skills supports one architecture, Linux ARM64. This is referred to as embedded (ARM64) throughout this documentation.
For more information and questions, visit the
.
Prerequisites
Before using Riva Speech AI, ensure you meet the following prerequisites:
You have access and are logged into NVIDIA NGC. For step-by-step instructions, refer to the
.
You have access to an NVIDIA Jetson Thor. For more information, refer to the
.
You have installed NVIDIA JetPack™ version 7.x on the Jetson platform. For more information, refer to the
.
You have ~32 GB free disk space on Jetson as required by the default containers and models. If you are deploying any Riva model intermediate representation (RMIR) models, the additional disk space required is ~32 GB plus the size of the RMIR models.
You have enabled the following power modes on the Jetson platform. These modes activate all CPU cores and clock the CPU/GPU at maximum frequency for achieving the best performance.
sudonvpmodel-m0(JetsonThor,modeMAXN)
You have set the default runtime to nvidia on the Jetson platform by adding the following line in the /etc/docker/daemon.json file. Restart the Docker service using sudosystemctlrestartdocker after editing the file.
"default-runtime":"nvidia"
Obtain a free trial license to install NVIDIA Riva. For more information, refer to the
.
Service Ports
The Riva Speech server listens on the following ports by default:
Port
Protocol
Used For
50051
gRPC
Riva gRPC APIs (used by Python clients such as transcribe_file.py and talk.py)
9001
HTTP/REST
Riva HTTP REST endpoints (for example, /v1/audio/synthesize, /v1/audio/transcriptions)
9000
WebSocket
Riva realtime WebSocket endpoints (for example, /v1/realtime?intent=synthesize, /v1/realtime?intent=transcription)
Configure the port numbers in config.sh of the Quick Start scripts using these fields:
riva_speech_api_port="50051"riva_speech_api_http_port="9001"riva_speech_api_realtime_port="9000"Deployment Guide
There is one push-button deployment option to deploy Riva Speech AI, which uses
available from the NGC catalog:
Local Docker: You can use the Quick Start scripts to set up a local workstation and deploy the Riva services using Docker. Continue with this guide to use the Quick Start scripts.
In addition to using pretrained models, Riva Speech AI can run with fine-tune custom models using
. Refer to the
section for details regarding the advanced option to create a model repository with NVIDIA NeMo.
For detailed instructions on deploying and using specific Riva services, refer to the following quick start guides.