An open API service indexing awesome lists of open source software.

https://github.com/interactivetech/riva-audio-agent

End to end demo showcasing how to deploy and run an end to end conversational agent in HPE's Private Cloud AI Environment
https://github.com/interactivetech/riva-audio-agent

Last synced: 12 months ago
JSON representation

End to end demo showcasing how to deploy and run an end to end conversational agent in HPE's Private Cloud AI Environment

Awesome Lists containing this project

README

          

# End-to-End Audio Conversational Agent in HPE's Private Cloud AI (PCAI) Platform

This repo guides you through:

1. **Docker** → build & push two images (WebSocket server & Gradio UI)
2. **Kubernetes Dev** → a quick `kubectl apply` to test locally
3. **Helm** → package & import into HPE Private Cloud AI (PCAI)

---

## Pre-requisites

Deploy the following models in your HPE Machine Learning Inference Software Tool:

- **magpie-tts-multilingual**
https://build.nvidia.com/nvidia/magpie-tts-multilingual/api
- **parakeet-ctc-1.1b-asr**
https://build.nvidia.com/nvidia/parakeet-ctc-1_1b-asr/modelcard
- **meta-llama/Llama-3.2-1B**
https://huggingface.co/meta-llama/Llama-3.2-1B

---

## Quickstart

1. **Docker**
```bash
cd docker
bash build.sh
```

See docker/README.md.

2. Kubernetes Dev
```bash
cd dev-k8s-deployment
kubectl apply -f pod-svc-vs.yaml
```

3. Helm & PCAI
```bash
cd helm
helm package .
```

Import audio-pipeline-chart-0.1.0.tgz via PCAI’s Tools & Frameworks → Import Framework.
See helm/README.md.

Now you’re all set to run your end-to-end audio AI pipeline on HPE Private Cloud AI!