GraalPy: Package Compatibility
GraalPy is compatible with many packages for Data Science and Machine Learning, including the popular PyTorch, NumPy, and Huggingface Transformers. To try a package, pick any version and only if you run into problems, consult our table below to see if there is a version that may work better.GraalPy 24.2
GraalPy 24.1
Numeric Computing
We test NumPy across multiple versions and know of multiple deployments where it brings numeric computing to Java.
Data Processing
Thanks to Arrow, Pandas on GraalPy can run multi-threaded while avoiding unneccessary data copies.
Models for any Task
The Huggingface transformers library works on GraalPy with its huge library of language, vision, and audio models.
Training and Inference
Train models and run inference on GraalPy with PyTorch, taking full advantage of the latest techniques and accellerator hardware.
Agentic Workflows
With Autogen and GraalPy you can write agentic workflows and use Java code to create tools for AI Agents.
To ensure GraalPy is compatible with common Python packages, the GraalPy team conducts compatibility testing and creates scripts to build and patch many of the top packages on PyPI plus some more that are of special interest to us, including libraries and frameworks such as NumPy, Pandas, and Django.
Compatibility testing ensures that developers can leverage GraalPy's powerful capabilities in their existing applications. It also enables developers to use GraalPy to create more efficient and productive applications in the areas of machine learning, data analysis, and web development using their familiar Python toolsets.
Many more Python packages than are on this list work on GraalPy. If there is a package you are interested in that you cannot find here, chances are that it might just work. If it does not, please reach out to us on GitHub
Compatible: loading...
Currently Untested: loading...
Currently
Incompatible: loading...
Not
Supported: loading...
Python Packages
| Name | Version | Notes |
|---|