joblib/joblib

Health v2 · maintenance only

Adoption not part of the score

4388 stars · 473 forks observed · 2026-08-28

What it is AI-extracted, prompt v1, taxonomy v1, 2026-08-29, confidence not recorded

Joblib is a Python library for lightweight pipelining: running Python functions in parallel with multiprocessing/threading backends and memoizing their results to disk. It also provides efficient serialization of large numpy arrays and optional Dask distributed execution.

Use cases

• parallelize a python function across cpu cores
• cache function results to disk in python
• memoize expensive computations
• dump and load large numpy arrays efficiently
• distribute python function execution with dask
• speed up embarrassingly parallel loops in python

When to choose

• you need simple parallel execution of Python functions without rewriting code
• you want transparent on-disk caching of function outputs
• you work with large numpy arrays and need fast persistence
• you want a lightweight alternative to full task-queue systems

When to avoid

• you need distributed task queues across many machines with a broker like Celery provides
• you need GPU or async-native parallelism
• your workload is I/O-bound network tasks better served by asyncio

Facets

library · maturity stable

caching serialization concurrency developer-tools data-science performance python cross-platform parallel-computing memoization multiprocessing numpy-serialization pipelining

Member repositories

For agents

markdown · JSON · MCP: product_card(name="joblib/joblib")