Metadocs Tutorials Repository

This repository contains tutorial code and resources linked to the Metadocs blog. Each folder corresponds to a specific tutorial, providing hands-on examples and insights into deploying advanced AI applications.

Folders

RAG-pipeline-langchain-openai

This folder contains resources for deploying a Retrieval-Augmented Generation (RAG) application using Langchain, Streamlit, and OpenAI.

Article: Deploy a RAG application with Langchain, Streamlit, and OpenAI in 10 minutes

Prerequisites:

• Python environment with pipenv
• OpenAI API key
• Basic understanding of LLMs, RAG pipelines, embeddings, and vector stores

Key Concepts Covered:

• Building a RAG pipeline
• Utilizing Langchain for application development
• Implementing Streamlit for web application deployment
• Handling embeddings and vector stores for data retrieval

deploy-llm-sagemaker-endpoint

This folder contains the necessary code and instructions to deploy an LLM using AWS Sagemaker and Hugging Face.

Article: Deploy a LLM on AWS in 5 minutes

Prerequisites:

• An AWS account with admin permissions
• AWS CLI configured on your machine
• Familiarity with Jupyter Lab, AWS Sagemaker, and Hugging Face's TGI
• Basic knowledge of Python and virtual environments

Key Concepts Covered:

• Setting up a Jupyter Lab environment with AWS Sagemaker
• Deploying Mistral 7B model on a Sagemaker endpoint
• Interacting with the deployed model through Jupyter notebook
• Properly deleting the Sagemaker endpoint to avoid additional charges