Metadata-Version: 2.4
Name: keras-neural-processes
Version: 0.1.0
Summary: Neural processes (CNP, NP, ANP) in Keras with ragged-tensor support for sparse, irregular time series
Author-email: Siddharth Chaini <sidchaini@gmail.com>
Project-URL: Repository, https://github.com/sidchaini/keras-neural-processes
Classifier: Development Status :: 4 - Beta
Classifier: Intended Audience :: Developers
Classifier: Intended Audience :: Education
Classifier: Intended Audience :: Science/Research
Classifier: License :: OSI Approved :: GNU General Public License v3 (GPLv3)
Classifier: Programming Language :: Python :: 3.10
Classifier: Programming Language :: Python :: 3.11
Classifier: Programming Language :: Python :: 3.12
Classifier: Programming Language :: Python :: 3.13
Classifier: Operating System :: OS Independent
Requires-Python: >=3.10
Description-Content-Type: text/markdown
Requires-Dist: numpy>=1.25.2
Requires-Dist: pandas>=2.0.3
Requires-Dist: scipy>=1.10
Requires-Dist: keras>=3
Requires-Dist: tensorflow
Requires-Dist: tensorflow-probability

# keras-neural-processes (`knp`)

Neural processes in Keras for sparse, irregular, variable-length time series (ragged tensors). 


We have implemented three model flavors:

| Class | Paper | Aggregation | Latent path |
|-------|-------|-------------|-------------|
| `CNP` | [(Garnelo et al. 2018a)](https://arxiv.org/pdf/1807.01613) | mean | no |
| `NP`  | [(Garnelo et al. 2018b)](https://arxiv.org/pdf/1807.01622) | mean | yes |
| `ANP` | [(Kim et al. 2019)](https://arxiv.org/pdf/1901.05761) | self + cross attention | yes |

For more details about this implementation, please refer to [(Chaini et. al. 2026)](https://arxiv.org/pdf/2605.27527). For the repository on application to astronomical light curves, refer to [NightLANP](https://github.com/sidchaini/NightLANP).
