Metadata-Version: 2.4
Name: opake-patches
Version: 0.16.1
Summary: Opake patches: PyTorch checkpoint shim and Hugging Face compatibility/kernel patches
Author: Opake Contributors
License-Expression: Apache-2.0
Keywords: differential-privacy,huggingface,transformers,pytorch
Classifier: Development Status :: 3 - Alpha
Classifier: Intended Audience :: Developers
Classifier: Intended Audience :: Science/Research
Classifier: Programming Language :: Python :: 3
Classifier: Programming Language :: Python :: 3.11
Classifier: Programming Language :: Python :: 3.12
Classifier: Programming Language :: Python :: 3.13
Classifier: Topic :: Scientific/Engineering :: Artificial Intelligence
Requires-Python: <3.14,>=3.11
Description-Content-Type: text/markdown
License-File: LICENSE
License-File: NOTICE
Requires-Dist: torch>=2.9.0
Requires-Dist: opake-engine==0.16.1
Provides-Extra: transformers
Requires-Dist: transformers<6,>=5.0.0; extra == "transformers"
Requires-Dist: numpy>=1.24; extra == "transformers"
Requires-Dist: peft>=0.13.0; extra == "transformers"
Dynamic: license-file

# opake-patches

Unified package for Opake patches, providing:

1. PyTorch checkpoint shim (`opake.patches.torch.runtime`)
2. Hugging Face runtime compatibility patches (`opake.patches.transformers.runtime`)
3. Hugging Face model/component patching (`opake.patches.transformers.models`, `opake.patches.transformers.components`)
4. Explicit PEFT/LoRA model patching (`opake.patches.peft`)

This package is pulled in by `opake` and should be consumed through the root
`opake` install surface.
