Master Vector Database - ChromaDB

Discover Chroma, an open source vector database with built-in embedding capabilities. Learn to create collections, add documents with metadata and IDs, and perform semantic search using embeddings and LLMs.

Explore methods on collections in ChromaDB, including counting items, retrieving documents, adding or upserting items, and viewing embeddings and metadata for each collection.

Explore storing movie character data in ChromaDB by creating and managing collections, adding documents with embeddings, and switching distance functions from L2 to cosine for efficient vector searches.

Learn how to add document associated embeddings in ChromaDB by creating a collection, attaching documents with embeddings and metadata, and querying by text or id using cosine distance.

Learn to use where filters to query a vector database, retrieve items by IDs, and perform updates, upserts, and deletions within collections, with examples like Morpheus and Agent Smith.

Load articles into chroma db, embed with OpenAI, and perform QA on multiple documents using LangChain and the GPT-3.5 turbo API in Colab.

Build a retrieval QA workflow with ChromaDB and LangChain, using OpenAI embeddings and GPT-3.5 Turbo to answer across multiple documents with controlled k and return source documents.

Build a ChromaDB stock collection by importing Wikipedia data and creating the collection with a default embedding, then query with filters by sector, location, or subsector.

Query stock companies in ChromaDB using text, a query vector, and a where filter on metadata to return matching metadata, documents, and distances.

Learn to store a pet knowledge base in ChromaDB using parquet-backed persistence, load text and PDF data, and apply embedding models for semantic search and filters.

Master semantic searches in chroma db using query embedding and where clause filters to retrieve relevant pet data, switching between default and paraphrase mini lm l3 embeddings.

Set up a chroma db vector store, load images with an image loader, and index embeddings for multimodal image search using open clip.

Configure and load the open clip embedding function, add images to the ChromaDB vector store, and perform multimodal image search using embedding queries and distance metrics.

Leveraging a multimodal model, this lecture demonstrates adding labeled food images to a ChromaDB vector store, querying by content such as carrot or broccoli, and refining results with metadata.

Attach metadata to images in ChromaDB to enhance multimodal image queries. Use category metadata and a where clause to filter animal and food results in the vector database.

Store a fed speech transcript in ChromaDB, then use LangChain and OpenAI to perform quick Q&A over the pdf-based knowledge store with a persistent vector store.

Learn to summarize a pdf by loading the document, extracting text, storing it in ChromaDB, and using OpenAI embeddings with LangChain to run a summarize chain.

Build a chat-based retrieval augmented generation system using ChromaDB and GPT-3.5 on Wikipedia content sources.

Explore how to use the Google Gemini Pro embedding model with ChromaDB to embed restaurant menu items, store documents, metadata, and IDs, and query the vector store with filters.