pychop
Python library for simulating low precision floating-point arithmetic
Python library for simulating low precision floating-point arithmetic
To install this package, run one of the following:
Pychop is a Python library designed to facilitate numerical experimentation with low-precision arithmetic. It supports customizable floating-point formats and a comprehensive set of rounding modes, integrating seamlessly with NumPy, PyTorch, and JAX for efficient low-precision emulation in deep learning and scientific computations.
Summary
Python library for simulating low precision floating-point arithmetic
Last Updated
Mar 15, 2026 at 21:56
License
MIT
Total Downloads
1.3K
Version Downloads
18
Supported Platforms
GitHub Repository
https://github.com/inexascale/pychopDocumentation
https://pychop.readthedocs.io/en/latest/?badge=latest