Hugging Face LLMs with SageMaker - RAG with Pinecone

Learn how to build Large Language Model (LLM) and Retrieval Augmented Generation (RAG) pipelines using open-source models from Hugging Face deployed on AWS SageMaker in this comprehensive video tutorial. Explore the implementation of semantic search using the MiniLM sentence transformer with Pinecone. Discover the process of deploying Hugging Face LLMs on SageMaker, generating LLM responses with context, and understanding the benefits of Retrieval Augmented Generation. Follow along as the instructor demonstrates deploying the MiniLM embedding model, creating context embeddings, and setting up a Pinecone vector index using the SageMaker FAQs dataset. Gain practical insights into making queries in Pinecone and implementing RAG for improved AI-powered applications. The tutorial also covers essential steps for managing and deleting running instances to optimize resource usage.

Open Source LLMs on AWS SageMaker
Open Source RAG Pipeline
Deploying Hugging Face LLM on SageMaker
LLM Responses with Context
Why Retrieval Augmented Generation
Deploying our MiniLM Embedding Model
Creating the Context Embeddings
Downloading the SageMaker FAQs Dataset
Creating the Pinecone Vector Index
Making Queries in Pinecone
Implementing Retrieval Augmented Generation
Deleting our Running Instances