liveBook · Manning

Pinecone

As explained in LLM Customization and Fine-Tuning

A managed vector database platform used here as an example of HNSW-based production vector search.

Pinecone is a managed vector database used for production vector search. A vector database adds an operational layer around an index, including a query API, metadata filters, persistence, replication, multi-tenancy, and observability. section 4.5.3 section 4.5.3

Choose the search layer to match the corpus and operating requirements:

  • Use a NumPy array for a prototype with tens of thousands of chunks, or an in-process index library such as FAISS for hundreds of thousands. section 4.5.3
  • Consider a vector database such as Pinecone for multi-tenant production with millions of chunks and audit logging. section 4.5.3

How LLM Customization and Fine-Tuning covers it

How deep it goes
Defines vector databases as an operational layer around indexes and identifies Pinecone as managed, without detailing its configuration.
Examples
Uses a multi-tenant production corpus with millions of chunks and audit logging as the scale case for choosing a vector database.
What you can do after
Evaluate whether corpus scale and production requirements warrant a managed vector database rather than an in-process index.
How it connects
Connects Pinecone to vector indexes, approximate nearest-neighbor search, and HNSW, alongside Weaviate, Qdrant, and pgvector.

Also explained in

Compare how 4 books explain Pinecone →

Related concepts

This explainer was generated by Manning's knowledge graph from the text of LLM Customization and Fine-Tuning. The section links point to the parts of the book it draws on.