# google/ml_collections

ML Collections is a library of Python Collections designed for ML use cases.

Repository: https://github.com/google/ml_collections
Canonical: https://ross.abutalabs.com/products/ml_collections
Homepage: https://ml-collections.readthedocs.io/
Language: Python
License: Apache-2.0
License Family: permissive
Last push: 2026-07-07T15:52:22+00:00

## Health v2 (maintenance only)
Score: 67/100 (v2, computed 2026-09-02T17:46:02.011165+00:00)
- activity 91, release rhythm 16, longevity 100
- inputs: {"age_days": 2204, "days_push": 57, "days_rel": 503, "gap_med": 156, "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 1036, forks 49 (observed 2026-08-28T04:03:19.146628+00:00)

## What it is
ML Collections is a Python library providing dict-like data structures (ConfigDict and FrozenConfigDict) for managing ML experiment and model configurations. It offers dot-based nested access, type safety, immutability, lazy computation, and YAML-readable printing.

## Use cases
- manage hyperparameter configs for ML experiments
- define model configuration files with dot access
- create immutable frozen configs for reproducible runs
- avoid typos in nested config dictionaries
- share experiment configs as keyword arguments

## When to choose
- you need structured, type-safe configuration for ML experiments
- you want immutable, hashable config objects
- you work in Python ML research pipelines (e.g., JAX/Flax ecosystems)

## When to avoid
- you need a general-purpose config format like YAML or TOML files alone
- your project is not Python-based
- you need runtime-editable configs with no locking semantics

## Facets
- artifact type: library
- maturity: stable
- function: configuration-management, developer-tools
- domain: machine-learning, developer-tools, deep-learning
- platform: python
- tags: configdict, experiment-configuration, hyperparameters, dot-access, type-safety

## Member repositories
- google/ml_collections (main) score 67

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:03:19.146628+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 2026-08-30T07:04:40.361743+00:00, confidence not recorded.
  - readme: https://github.com/google/ml_collections (fetched 2026-08-28T04:03:19.146628+00:00, sha 84fc111467bb)
  - registry_pypi: https://pypi.org/pypi/ml_collections/json (fetched 2026-08-29T13:06:15.730637+00:00, sha 17c29907e86d)
- Data as of 2026-08-30T08:39:29.467469+00:00.