clu
Set of libraries for ML training loops in JAX.
Decision gist · record as of 2026-08-14
Yes, if you are writing training loops in JAX or Flax. CLU is actively maintained, has no security vulnerabilities, low install friction, and is widely used in the JAX ecosystem. The permissive Apache 2.0 license poses no restrictions. Verify that CLU's feature set matches your specific checkpoint and metrics needs before committing.AI-flagged interpretation of the facts on this page — verify before relying
Before you install
- Requires JAX and jaxlib installed; CLU is a library for JAX-based training, not a standalone tool.
- Low friction: pure Python wheel with no compiled dependencies.
- Active maintenance—last commit 2026-07-07, though the latest release was 2024-04-10, suggesting the package is stable but not under rapid iteration.
License · maintenance · safety
Apache 2.0 (permissive) — Apache 2.0 is permissive; you can use, modify, and distribute CLU freely in commercial and private projects, provided you retain the license notice.
last release 2024-04-10 (856 days) · last repo commit 2026-07-07 · 370 stars
0 known vulnerabilities (OSV.dev, 2026-08-14) · 573,809 downloads/mo, #5,944 on PyPI
Alternatives
Verify before relying
pip install clu
import clu
from clu import metrics
# Use CLU metrics and utilities in your JAX training loop
metric_collection = metrics.Collection()- Specific metrics types and checkpoint formats supported by CLU beyond what the excerpt describes.
- Whether CLU is actively accepting bug fixes or is in maintenance-only mode given the time since last release.
What it is and what it does
CLU is a set of libraries built on top of JAX for structuring machine learning training loops. It abstracts common tasks—metrics aggregation, checkpointing, and loop orchestration—into reusable components so that training code stays readable and concise without sacrificing the flexibility needed for research. The package depends on JAX, Flax, ml-collections, and several utility libraries (absl-py, etils, numpy, wrapt, packaging, typing-extensions) to provide a cohesive toolkit.
The library is designed for researchers and practitioners writing custom training loops in JAX who want to avoid boilerplate while keeping full control over loop logic. It is actively maintained (last commit 2026-07-07) and widely adopted in the JAX and machine learning community. The project is not currently accepting external contributions but can be forked for custom extensions.
Use it for
- Collecting and aggregating metrics (loss, accuracy, etc.) during JAX training without writing custom metric logic.
- Managing checkpoints and model state during long-running training jobs in JAX.
- Structuring a custom training loop in Flax or raw JAX while reusing common patterns from CLU.
- Prototyping ML experiments in JAX with minimal boilerplate for loop orchestration and logging.
Worth the install?
AI-flagged interpretation of the facts on this page. Verify before relying on it.
Yes, if you are writing training loops in JAX or Flax.
CLU is actively maintained, has no security vulnerabilities, low install friction, and is widely used in the JAX ecosystem. The permissive Apache 2.0 license poses no restrictions. Verify that CLU's feature set matches your specific checkpoint and metrics needs before committing.
Install
clu on PyPI
Before you install
Low friction: pure Python wheel with no compiled dependencies. Active maintenance—last commit 2026-07-07, though the latest release was 2024-04-10, suggesting the package is stable but not under rapid iteration.
Requires JAX and jaxlib installed; CLU is a library for JAX-based training, not a standalone tool.
License in practice
Apache 2.0 is permissive; you can use, modify, and distribute CLU freely in commercial and private projects, provided you retain the license notice.
Quickstart
pip install clu
import clu
from clu import metrics
# Use CLU metrics and utilities in your JAX training loop
metric_collection = metrics.Collection()Verify before relying
- Specific metrics types and checkpoint formats supported by CLU beyond what the excerpt describes.
- Whether CLU is actively accepting bug fixes or is in maintenance-only mode given the time since last release.
Package facts
| License | Apache 2.0 permissive |
| Python support | Not specified |
| Install friction | Low. Pure-Python wheel |
| Runtime dependencies | 10 packagesabsl-pyetilsflaxjaxjaxlibml-collectionsnumpypackagingtyping-extensionswrapt |
| Maintenance | Actively maintained 856 days since the last release |
| Last repo commit | |
| First released | |
| Downloads | 573,809 / month, #5,944 on PyPI 30-day window, as of 2026-08-14 |
| Known vulnerabilities | None known OSV.dev, checked 2026-08-14 |
| Classifiers | Development Status :: 4 - BetaIntended Audience :: DevelopersIntended Audience :: Science/ResearchLicense :: OSI Approved :: Apache Software LicenseProgramming Language :: PythonTopic :: Scientific/Engineering :: Artificial Intelligence |
Evidence: clu-0.0.12-py3-none-any.whl
Tags
Let your AI agent find packages like this
Example. Real query, live index.
You found this page by searching. An agent finds it by wishing: SkillFed indexes 14,416 PyPI packages by what they can do, searchable in plain language.
wish › “JAX training loop utilities”
- cluCLU provides utilities and abstractions for writing machine learning…
- equinoxEquinox provides neural network and model building on top of JAX with…
- augmaxAugmax is a JAX-based image data augmentation framework that chains…
Give your agent the search over MCP, or paste the wish link into any chat.
More Artificial Intelligence packages
LiteLLM provides a unified Python interface to call 100+ LLM providers (OpenAI, Anthropic, Gemini, Bedrock, Azure, and others) using OpenAI-compatible API format, available as both a Python SDK and a self-hosted AI Gateway proxy server.
Install it if you need to work with multiple LLM providers or want to centralize LLM routing in your organization.
Client library and CLI tool for downloading, uploading, and managing models, datasets, and repositories on the Hugging Face Hub platform.
Install it if you work with Hugging Face Hub models or datasets.
LangChain provides a framework for building agents and LLM-powered applications by composing language models, tools, and memory through a unified API that abstracts over multiple model providers.
hf-xet provides chunk-based deduplication and efficient file transfer for the Hugging Face Hub, enabling faster uploads and downloads of large files with local disk caching.
Tokenizers converts raw text into token sequences for NLP models, with support for training custom vocabularies and using pre-built tokenizers (BPE, WordPiece) optimized for speed via Rust.
Transformers provides a unified framework for loading, fine-tuning, and running state-of-the-art pretrained models across text, vision, audio, video, and multimodal tasks using PyTorch, JAX, or TensorFlow.
Install it if you need to run or train any transformer-based model for NLP, vision, audio, or multimodal tasks.
See also flax · orbax-checkpoint · rax · dm-haiku · drjax · google-tunix · jmp · aqtp · google-metrax · optax