This repository consists of gpu bootcamp material for RIVA. The RIVA frameworks of libraries, APIs and models allows you to build, optimize and deploy voice and speech centric applications and models based entirely on GPUs. In this series you can access RIVA learning resources in the form of labs. The modules covered in this Bootcamp are SpeechToText, Intent Slot Classification, Named Entity Recognition, Question-Answering and Challenge.
To run this tutorial you will need a machine with NVIDIA GPU.
-
Install the latest Docker .
-
The base containers required for the lab requires users to create a NGC account and generate an API key (https://docs.nvidia.com/ngc/ngc-catalog-user-guide/index.html#registering-activating-ngc-account)
To start with, you will have to pull the RIVA Docker container
ngc registry resource download-version "nvidia/riva/riva_quickstart:1.4.0-beta"
cp *.sh riva_quickstart_v1.4.0-beta
cd riva_quickstart_v1.4.0-beta
If running with multiple users/clients, please choose a UNIQUE Riva server port number for each user (default is 50051)
bash riva_init_new.sh
bash riva_start.sh
bash riva_start_bootcamp.sh
jupyter notebook --ip=0.0.0.0 --allow-root --notebook-dir=/bootcamp --no-browser --port=8888 --NotebookApp.token=''
Then, open the jupyter notebook in browser: http://localhost:8888
Start working on the lab by clicking on the Start_Here_RIVA.ipynb notebook.
Q. Cannot run RIVA scripts
A. The Riva platform requires Docker to be actively running. In case you are running rootless-docker you should ensure that the Docker daemon is active and has sufficient permissions
Q. Invalid API key
A. Pulling pretrained model containers from NGC requires a valid API key. To get this key, please login to you NGC account and generate a new valid API key and re initialize the Riva platform