Skip to main content
NetApp Solutions

Deploying support center sentiment analysis

Contributors kevin-hoke
PDFs

Deploying the solution involves the following components:

  1. NetApp DataOps Toolkit

  2. NGC Configuration

  3. NVIDIA RIVA Server

  4. NVIDIA TAO Toolkit

  5. Export TAO models to RIVA

To perform deployment, complete the following steps:

NetApp DataOps Toolkit: Support center sentiment analysis

To use the NetApp DataOps Toolkit, complete the following steps:

  1. Pip install the toolkit.

    python3 -m pip install netapp-dataops-traditional

  2. Configure the data management

    netapp_dataops_cli.py config

NGC configuration: Support center sentiment analysis

To set up NVIDIA NGC, complete the following steps:

  1. Download the NGC.

    wget -O ngccli_linux.zip https://ngc.nvidia.com/downloads/ngccli_linux.zip && unzip -o ngccli_linux.zip && chmod u+x ngc

  2. Add your current directory to path.

    echo "export PATH=\"\$PATH:$(pwd)\"" >> ~/.bash_profile && source ~/.bash_profile

  3. You must configure NGC CLI for your use so that you can run the commands. Enter the following command, including your API key when prompted.

    ngc config set

For operating systems that are not Linux-based, visit here.

NVIDIA RIVA server: Support center sentiment analysis

To set up NVIDIA RIVA, complete the following steps:

  1. Download the RIVA files from NGC.

    ngc registry resource download-version nvidia/riva/riva_quickstart:1.4.0-beta

  2. Initialize the RIVA setup (riva_init.sh).

  3. Start the RIVA server (riva_start.sh).

  4. Start the RIVA client (riva_start_client.sh).

  5. Within the RIVA client, install the audio processing library ( FFMPEG)

    apt-get install ffmpeg

  6. Start the Jupyter server.

  7. Run the RIVA Inference Pipeline Notebook.

NVIDIA TAO Toolkit: Support center sentiment analysis

To set up NVIDIA TAO Toolkit, complete the following steps:

  1. Prepare and activate a virtual environment for TAO Toolkit.

  2. Install the required packages.

  3. Manually pull the image used during training and fine-tuning.

    docker pull nvcr.io/nvidia/tao/tao-toolkit-pyt:v3.21.08-py3

  4. Start the Jupyter server.

  5. Run the TAO Fine-Tuning Notebook.

Export TAO models to RIVA: Support center sentiment analysis

To use TAO Toolkit models in RIVA, complete the following steps:

  1. Save models within the TAO Fine-Tuning Notebook.

  2. Copy TAO trained models to the RIVA model directory.

  3. Start the RIVA server (riva_start.sh).

Deployment roadblocks

Here are a few things to keep in mind as you develop your own solution:

  • The NetApp DataOps Toolkit is installed first to ensure that the data storage system runs optimally.

  • NVIDIA NGC must be installed before anything else because it authenticates the downloading of images and models.

  • RIVA must be installed before the TAO Toolkit. The RIVA installation configures the docker daemon to pull images as needed.

  • DGX and docker must have internet access to download the models.