Cerebras Inference API Demos

Welcome to the Cerebras Inference API demo repository! This repository contains various examples showcasing the power of the Cerebras Wafer-Scale Engines and CS-3 systems for AI model inference.

🚀 Introduction

The Cerebras Inference API offers developers a low-latency solution for AI model inference powered by Cerebras Wafer-Scale Engines and CS-3 systems. We invite developers to explore the new possibilities that our high-speed inferencing solution unlocks.

The Cerebras Inference API provides access to models such as OpenAI's GPT-OSS, Meta's Llama family of models, and Alibaba's Qwen models. For the full details of supported models, see the supported models documentation.

📚 Resources

• Play with our live chatbot demo
• Experiment with our inference solution in the playground
• Explore our API reference documentation

📁 Projects Overview

This repository contains multiple example projects, each demonstrating different capabilities of the Cerebras Inference API. Each project is located in its own folder and contains a detailed README.

🔗 Example Projects

• Getting Started with Cerebras Inference API

  • Learn how to get started with the Cerebras Inference API for your AI projects.

• Conversational Memory with Langchain

  • Explore how to build conversational memory for LLMs using Langchain.

• RAG with Pinecone + Docker

  • Implement Retrieval-Augmented Generation (RAG) using Pinecone and Docker.

• RAG with Weaviate + HuggingFace

  • Implement Retrieval-Augmented Generation (RAG) using Weaviate and HuggingFace.

• Getting Started with Cerebras + Streamlit

  • Learn how to integrate Cerebras with Streamlit to build interactive applications.

• AI Agentic Workflow with LlamaIndex

  • Build an AI agentic workflow using LlamaIndex.

• AI Agentic Workflow with Langchain

  • Build an AI agentic workflow using Langchain.

• Multi AI Agentic Workflow

  • Create a multi-agentic AI workflow with Langgraph and LangSmith.

• Payload Compression (msgpack + gzip)

  • Minimal example + benchmark showing msgpack+gzip request-body compression on /v1/chat/completions and the TTFT / E2E latency gain on a ~30k-token prompt.

---

🌟 Getting Started

To explore each project, simply navigate to the corresponding folder and follow the instructions in the README. Happy coding!

🛠️ Requirements

• Python 3.7+
• Docker (for RAG examples)
• Streamlit (for Cerebras + Streamlit example)
• Other dependencies as noted in each project’s README.

📝 License

This project is licensed under the MIT License - see the LICENSE file for details.

👥 Contributors

We welcome contributions! Feel free to submit a pull request or open an issue.

---

© 2024 Cerebras Systems