faiss

Name: faiss
Availability: InStock
Author: aissenliu

Perform vector similarity search across billions of vectors using FAISS index types.

Updated Apr 27, 2026

Search Tags:#indexing #gpu #similarity #knn #vector-search #machine-learning #faiss

Ready to run

npx skills add https://github.com/AissenLiu/EasyHermes --skill faiss-aissenliu

💬 Prompt for ChatGPT / Claude Chat▼

Please help me install this Agent Skill.
Skill: faiss
Source: https://github.com/AissenLiu/EasyHermes/tree/main/hermes-agent/optional-skills/mlops/faiss
Command: npx skills add https://github.com/AissenLiu/EasyHermes --skill faiss-aissenliu

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill requires faiss-cpu, faiss-gpu, numpy, and includes references (resource) components.

What problem does it solve?

FAISS provides fast, scalable vector similarity search for large-scale datasets, enabling near-real-time retrieval of similar vectors across billions of entries.

Core Features & Use Cases

Multiple index types: Flat, IVF, HNSW, and PQ for exact and approximate search.
Hardware acceleration: GPU support for large-scale deployments and faster queries.
Indices persistence: Save and load trained indices for offline and online use.
Integration readiness: Compatible with LangChain and other ML tooling for end-to-end retrieval pipelines.
Common use cases: Nearest neighbor search for embeddings in NLP, recommendation, and retrieval-augmented generation.

Quick Start

Install FAISS (cpu or gpu), create an index, add vectors, and run a search to retrieve nearest neighbors.

Frequently Asked Questions about faiss

High-intent search queries and answers about installing and using this skill.

FAQPage Schema

How do I perform fast vector similarity search for billions of vectors?▼

Fast vector similarity search for billions of vectors is achieved by creating FAISS indices like IVF, HNSW, or PQ to retrieve nearest neighbors efficiently. This enables near-real-time retrieval across massive datasets for ML pipelines.

Can I run FAISS vector search on a GPU for faster queries?▼

Yes, FAISS vector search supports GPU execution to accelerate large-scale deployments and queries. By utilizing faiss-gpu, you can retrieve nearest neighbors significantly faster than CPU-only processing for massive datasets.

What is the best way to save and load trained indices for offline vector search?▼

The best way to manage trained indices for offline vector search is using FAISS index persistence features. You can save and load trained indices directly, enabling seamless integration between offline dataset processing and online retrieval workflows.

When should I use approximate nearest neighbor search instead of exact Flat indices?▼

Use approximate nearest neighbor search indices like IVF, HNSW, or PQ when scalability across billions of vectors is required. Exact Flat indices provide perfect accuracy but are too slow for large-scale machine learning pipelines.

Does vector similarity search integrate with LangChain and NLP embedding workflows?▼

Vector similarity search integrates directly with LangChain and NLP embedding workflows. FAISS acts as the retrieval backend, enabling end-to-end retrieval-augmented generation pipelines by efficiently querying dense text embeddings.

What do I need to install to start using FAISS for knn search?▼

To start using FAISS for knn search, you need to install faiss-cpu or faiss-gpu alongside numpy. These dependencies enable you to create indices, add vectors, and run nearest neighbor retrieval queries immediately.