{"categories":[{"label":"Artificial Intelligence","url":"https://skillfed.io/packages/category/scientific-engineering-artificial-intelligence/10"}],"enrichment":{"capability":"Jax-jumpy provides a unified interface that runs NumPy code as-is when JAX is unavailable, and automatically switches to JAX's accelerated backend when JAX is installed and conditions are met (JIT compilation or JAX inputs).","skillfed_tags":["backend-abstraction","jax-numpy-interop"],"use_cases":["Write RL environment wrappers that work with both NumPy and JAX backends without duplicating logic.","Debug numerical code in plain NumPy, then run it accelerated on GPU/TPU via JAX without code changes.","Build libraries that optionally use JAX acceleration when available, falling back gracefully to NumPy.","Implement array operations that automatically dispatch to the fastest available backend at runtime."],"what_it_does":"Jax-jumpy is a compatibility layer that lets you write code once and run it on either NumPy or JAX depending on what is installed and how the code is called. The core idea is to enable framework-agnostic libraries\u2014particularly reinforcement learning environments like Gymnasium and PettingZoo\u2014to work seamlessly with both standard NumPy arrays and JAX-accelerated versions without code duplication.\n\nWhen you call a jax-jumpy function, it checks whether JAX is available and whether the inputs are JAX arrays or the function is JIT-compiled; if so, it runs the JAX version for acceleration. Otherwise it falls back to NumPy. This design lets you debug with plain NumPy (which is easier to inspect and step through) and deploy with JAX (for hardware acceleration) using the same codebase. The package currently supports a subset of NumPy and JAX functions, with some advanced features like `vmap` and `scan` available only when JAX is installed.","worth_installing":"Yes, if you are building libraries or environments that need to support both NumPy and JAX users without code duplication. The low install friction, active maintenance, permissive license, and zero known vulnerabilities make it a safe choice. Not necessary if you are already committed to a single backend or do not need optional acceleration."},"id":"jax-jumpy","links":{"html":"https://skillfed.io/packages/jax-jumpy","md":"https://skillfed.io/packages/jax-jumpy.md","pypi":"https://pypi.org/project/jax-jumpy/"},"maintenance":{"status":"active"},"meta":{"latest_release":"2023-03-17","license_spdx":null,"license_treatment":"permissive","name":"jax-jumpy","python_support":"supports_current","summary":"Common backend for Jax or Numpy."},"popularity":{"monthly_downloads":96724,"position":13197,"tier":"top_15000"},"security":{"n_vulnerabilities":0},"version":"1.0.0"}