nwf-core 0.2.1

Core library for Neural Weight Fields (NWF) - charge, field, metrics, indices

nwf-core

Core library for Neural Weight Fields (NWF) - charge, field, metrics, indices, calibration.

Installation

pip install nwf-core
# With FAISS: pip install nwf-core[faiss]
# With VAE (torch): pip install nwf-core[torch]
# All: pip install nwf-core[all]

Quick start

import numpy as np
from nwf import Charge, Field, mahalanobis_symmetric

# Create charges
c1 = Charge(z=np.array([0.0, 0.0]), sigma=np.array([0.1, 0.1]))
c2 = Charge(z=np.array([1.0, 1.0]), sigma=np.array([0.1, 0.1]))

# Build field and search
field = Field()
field.add([c1, c2], labels=[0, 1])
distances, indices, labels = field.search(c1, k=2)

See documentation and nwf-vision for examples.

Components

• Charge: (z, sigma) - center and diagonal covariance
• Field: Container for charges with add/remove/search
• Metric: mahalanobis_symmetric, euclidean, cosine
• Index: BruteForceIndex, FAISSIndex (l2, cosine, ip; optional rerank with Mahalanobis)
• Calibration: AgreementRatio, PlattScaler
• Encoders: VAEEncoder (optional, requires torch)

Development

pip install -e ".[dev,docs]"
pre-commit install   # optional: run checks before commit
make lint            # black, isort, flake8
make format          # auto-format
make html            # build Sphinx docs -> docs/_build/html

License

MIT