opea-project/GenAIExamples

Generative AI Examples is a collection of GenAI examples such as ChatQnA, Copilot, which illustrate the pipeline capabilities of the Open Platform for Enterprise AI (OPEA) project.

What it solves

GenAIExamples provides a set of microservice-based reference implementations to help developers quickly deploy, test, and scale generative AI applications across diverse hardware platforms. It removes the friction of setting up complex GenAI pipelines by providing pre-configured examples for common enterprise use cases.

How it works

The project uses a modular architecture based on three core pillars:

  • GenAIComps: A library of microservice components (such as LLMs, embedding models, and rerankers) used to build specific applications.
  • GenAIInfra: A cloud-native suite that handles the containerization and deployment of these examples via Docker Compose or Kubernetes (using Helm or GMC).
  • GenAIEval: A performance measurement tool used to track throughput, latency, and accuracy across different hardware configurations.

Who it’s for

Developers and enterprise AI engineers who need to deploy GenAI applications on various hardware (including Intel Xeon, Gaudi, AMD EPYC/Instinct, and NVIDIA GPUs) and compare their performance.

Highlights

  • Multi-Hardware Support: Native compatibility with Gaudi, Xeon, AMD EPYC, AMD Instinct, and NVIDIA GPUs.
  • Diverse Use Cases: Includes ready-to-deploy examples for RAG-based ChatQnA, VisualQnA, Text-to-Image, Document Summarization, and Code Generation.
  • Cloud-Native Deployment: Supports flexible deployment via Python, Docker Compose, and Kubernetes.
  • Performance Benchmarking: Integrated evaluation tools to compare service performance across different hardware setups.