# tensorflow/datasets

TFDS is a collection of datasets ready to use with TensorFlow, Jax, ...

Repository: https://github.com/tensorflow/datasets
Canonical: https://ross.abutalabs.com/products/tensorflow-datasets
Homepage: https://www.tensorflow.org/datasets
Language: Python
License: Apache-2.0
License Family: permissive
Topics: tensorflow, machine-learning, data, datasets, numpy, jax, dataset
Last push: 2026-08-21T16:29:09+00:00

## Health v2 (maintenance only)
Score: 85/100 (v2, computed 2026-09-03T02:20:16.233290+00:00)
- activity 98, release rhythm 59, longevity 100
- inputs: {"age_days": 2914, "days_push": 12, "days_rel": 117, "gap_med": 133, "n_releases_24m": 4}
- flags: none
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 4581, forks 1590 (observed 2026-08-28T04:08:54.534836+00:00)

## What it is
TensorFlow Datasets (TFDS) is a curated collection of ready-to-use public datasets exposed as tf.data.Dataset objects for TensorFlow, JAX, and other Python ML frameworks. It handles downloading, preparation, versioning, and deterministic, high-performance input pipelines.

## Use cases
- load mnist for training a model
- get standard benchmark datasets for machine learning experiments
- build reproducible input pipelines in tensorflow or jax
- download and prepare public datasets with one line of code
- share a custom dataset in a versioned format with teammates
- convert a tf.data pipeline into a shareable dataset

## When to choose
- you need common ML datasets (vision, text, audio) with minimal setup
- you want deterministic, reproducible splits and ordering across users
- you train models in TensorFlow or JAX and want fast tf.data pipelines
- you need versioned, documented datasets for research citations

## When to avoid
- your data is proprietary or highly custom with no existing builder
- you work outside the Python ML ecosystem
- you need datasets for frameworks with incompatible formats and don't want tf.data

## Facets
- artifact type: dataset
- maturity: stable
- function: machine-learning, data-science, etl
- domain: machine-learning, data-science, deep-learning
- platform: python, cross-platform
- tags: tensorflow, jax, tfdata, datasets, numpy, input-pipelines

## Member repositories
- tensorflow/datasets (main) score 85

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:08:54.534836+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 2026-08-29T18:19:54.124091+00:00, confidence not recorded.
  - readme: https://github.com/tensorflow/datasets (fetched 2026-08-28T04:08:54.534836+00:00, sha 8ce81906e6d0)
  - homepage: https://www.tensorflow.org/datasets (fetched 2026-08-29T09:05:29.142027+00:00, sha 8d98a33eb27f)
  - site_page: https://www.tensorflow.org/install (fetched 2026-08-29T09:05:29.151236+00:00, sha 584a762da891)
  - site_page: https://www.tensorflow.org/tfx/api_docs (fetched 2026-08-29T09:05:29.154952+00:00, sha 6977825696fe)
  - site_page: https://www.tensorflow.org/about (fetched 2026-08-29T09:05:29.158109+00:00, sha 817250744d91)
  - site_page: https://www.tensorflow.org/about/case-studies (fetched 2026-08-29T09:05:29.159966+00:00, sha e6701029eec0)
  - site_page: https://www.tensorflow.org/datasets/api_docs/python/tfds/all_symbols (fetched 2026-08-29T09:05:29.161837+00:00, sha 2ddcdb431d39)
  - site_page: https://www.tensorflow.org/community/contribute (fetched 2026-08-29T09:05:29.156465+00:00, sha 77d394fa76e7)
  - site_page: https://www.tensorflow.org/datasets/format_specific_dataset_builders (fetched 2026-08-29T09:05:29.163540+00:00, sha 092a1ef85535)
- Data as of 2026-08-30T08:39:29.467469+00:00.