Generates and updates PyTorch-compliant docstrings using Sphinx formatting and official framework conventions.
This skill provides specialized guidance for writing professional-grade documentation for PyTorch functions and methods. It ensures strict adherence to the PyTorch framework's conventions, including the use of raw strings for LaTeX compatibility, Sphinx/reStructuredText (reST) formatting, and standardized sections for arguments, return types, and mathematical formulas. Whether you are contributing to the PyTorch core or building compatible libraries, this skill automates the creation of complex signatures and cross-references required for high-quality machine learning documentation.
Key Features
01Automates the creation of standardized Args, Returns, and Examples sections.
02Handles specialized patterns like C-bound functions, in-place variants, and aliases.
03Standardizes documentation using Sphinx and reStructuredText (reST) formatting.
0489 GitHub stars
05Enforces the use of raw strings to prevent LaTeX and math backslash errors.
06Supports mathematical equations and tensor shape documentation using LaTeX.
Use Cases
01Updating legacy machine learning codebases to match official PyTorch documentation styles.
02Generating cross-referenced documentation for third-party PyTorch-based libraries.
03Writing standardized documentation for new PyTorch neural network modules and functions.
