# tensorflow/tfx

TFX is an end-to-end platform for deploying production ML pipelines

Repository: https://github.com/tensorflow/tfx
Canonical: https://ross.abutalabs.com/products/tfx
Homepage: https://tensorflow.github.io/tfx/
Language: Python
License: Apache-2.0
License Family: permissive
Topics: tensorflow, machine-learning, apache-beam
Last push: 2026-08-17T22:09:07+00:00

## Health v2 (maintenance only)
Score: 91/100 (v2, computed 2026-09-03T02:20:16.233290+00:00)
- activity 98, release rhythm 77, longevity 100
- inputs: {"age_days": 2767, "days_push": 16, "days_rel": 76, "gap_med": 38, "n_releases_24m": 6}
- flags: none
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 2190, forks 724 (observed 2026-08-28T04:06:24.062452+00:00)

## What it is
TensorFlow Extended (TFX) is an end-to-end, Google-production-scale platform for building and deploying production machine learning pipelines based on TensorFlow. It provides a configuration framework of composable components (data validation, transformation, training, evaluation, serving) that can be orchestrated with Apache Airflow or Kubeflow Pipelines.

## Use cases
- deploy machine learning models to production pipelines
- orchestrate ml training pipelines with airflow or kubeflow
- validate and monitor training and serving data
- track ml metadata and experiment runs
- move tensorflow models from research to production
- build reproducible end-to-end ml workflows

## When to choose
- you use TensorFlow and need production-scale ML pipelines
- you need orchestration, data validation, and model serving in one platform
- you want ML metadata tracking and pipeline warmstarting

## When to avoid
- you use PyTorch or non-TensorFlow frameworks
- you only need lightweight experimentation or notebooks
- you want a simple single-model deployment without pipeline overhead

## Facets
- artifact type: framework
- maturity: stable
- function: machine-learning, etl, workflow-automation, data-science, monitoring
- domain: machine-learning, deep-learning
- platform: python, cloud
- tags: mlops, tensorflow, apache-beam, ml-pipelines, kubeflow, airflow, model-deployment, data-validation, data-engineering, pipelines, kubernetes, docker, linux

## Member repositories
- tensorflow/tfx (main) score 91

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:06:24.062452+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 2026-08-30T02:47:42.157339+00:00, confidence not recorded.
  - readme: https://github.com/tensorflow/tfx (fetched 2026-08-28T04:06:24.062452+00:00, sha 894c7968bdde)
  - homepage: https://tensorflow.github.io/tfx/ (fetched 2026-08-29T10:27:59.721108+00:00, sha 582884e17bc9)
  - registry_pypi: https://pypi.org/pypi/tfx/json (fetched 2026-08-29T10:27:59.730405+00:00, sha c07b2c4993e9)
- Data as of 2026-08-30T08:39:29.467469+00:00.