# google-parfait/tensorflow-federated

An open-source framework for machine learning and other computations on decentralized data.

Repository: https://github.com/google-parfait/tensorflow-federated
Canonical: https://ross.abutalabs.com/products/tensorflow-federated
Language: Python
License: Apache-2.0
License Family: permissive
Last push: 2026-08-26T03:29:01+00:00

## Health v2 (maintenance only)
Score: 79/100 (v2, computed 2026-09-03T02:20:16.233290+00:00)
- activity 99, release rhythm 40, longevity 100
- inputs: {"age_days": 2821, "days_push": 7, "days_rel": 706, "gap_med": 8, "n_releases_24m": 2}
- flags: none
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 2448, forks 606 (observed 2026-08-28T04:06:52.800812+00:00)

## What it is
TensorFlow Federated (TFF) is an open-source Python framework for machine learning and other computations on decentralized data. It provides high-level federated learning APIs built on TensorFlow plus a lower-level Federated Core for expressing novel federated algorithms, with a single-machine simulation runtime.

## Use cases
- train a model across many clients without collecting their data
- simulate federated averaging experiments on a single machine
- apply federated training and evaluation to existing TensorFlow models
- prototype novel federated learning algorithms
- compute aggregated analytics over decentralized datasets
- train mobile keyboard prediction models without uploading sensitive user data

## When to choose
- you need privacy-preserving training where data stays on client devices
- you are researching or prototyping federated learning algorithms
- your models are already built in TensorFlow
- you want a simulation runtime for federated experiments without real device infrastructure

## When to avoid
- you need production federated learning on real mobile devices rather than simulation
- your stack is PyTorch or another non-TensorFlow framework
- you need simple centralized training with no decentralization constraints
- you need lightweight analytics without learning a strongly-typed federated computation model

## Facets
- artifact type: framework
- maturity: active
- function: machine-learning, llm-training, simulation, sdk
- domain: machine-learning, artificial-intelligence, privacy, microservices
- platform: python, cross-platform
- tags: federated-learning, tensorflow, decentralized-data, privacy-preserving, simulation-runtime

## Member repositories
- google-parfait/tensorflow-federated (main) score 79

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:06:52.800812+00:00.
- Health v2: computed from the inputs above; adoption is never an input.
- Inferred fields (summary, facets, guidance): AI-extracted, prompt v1, taxonomy v1, on 2026-08-30T02:30:28.566387+00:00, confidence not recorded.
  - readme: https://github.com/google-parfait/tensorflow-federated (fetched 2026-08-28T04:06:52.800812+00:00, sha 07092f6931cd)
  - registry_pypi: https://pypi.org/pypi/tensorflow-federated/json (fetched 2026-08-29T10:11:51.122242+00:00, sha a79a3ec96586)
- Data as of 2026-08-30T08:39:29.467469+00:00.