facebookresearch/CrypTen

Health v2 · maintenance only

Flags: archived

Adoption not part of the score

1649 stars · 304 forks observed · 2026-08-28

What it is AI-extracted, prompt v1, taxonomy v1, 2026-08-30, confidence not recorded

CrypTen is a PyTorch-based library for privacy-preserving machine learning built on secure multiparty computation. It exposes a CrypTensor object that behaves like a regular PyTorch tensor, enabling encrypted training and inference with automatic differentiation.

Use cases

• train machine learning models on encrypted data
• run secure inference without revealing inputs
• perform secure multiparty computation with a PyTorch-like API
• research privacy-preserving machine learning techniques
• compute statistics across parties without sharing raw data
• train an SVM or neural network on encrypted tensors

When to choose

• you need SMPC-based privacy-preserving ML with a PyTorch feel
• you are an ML researcher exploring secure computation
• you want automatic differentiation over encrypted tensors
• you work on Linux or macOS with Python and optionally GPUs

When to avoid

• you need a production-ready secure ML system
• you need Windows support
• you need lightweight encryption rather than full multiparty computation
• you need actively maintained software with frequent releases

Facets

library · maturity maintenance

machine-learning cryptography security privacy machine-learning security privacy artificial-intelligence python secure-multiparty-computation pytorch privacy-preserving-ml research-framework encrypted-tensors linux macos gpu

Member repositories

For agents

markdown · JSON · MCP: product_card(name="facebookresearch/CrypTen")