alexshtf/torchcurves

Parametric differentiable curves with PyTorch for continuous embeddings, shape-restricted models, or KANs

47
/ 100
Emerging

This tool helps machine learning engineers and researchers build models that incorporate smoothly changing relationships, enforce specific data behaviors, or capture complex feature interactions. It takes numerical inputs and produces continuous, differentiable outputs that can represent embeddings for various categories, enforce monotonicity in predictions like auction win probabilities, or form the basis of advanced neural networks like Kolmogorov-Arnold Networks (KANs). It is designed for those who work with PyTorch and need fine-grained control over curve-based function approximation.

Available on PyPI.

Use this if you need to integrate flexible, learnable parametric curves (like B-splines or Legendre polynomials) directly into your PyTorch models for tasks like creating continuous embeddings, ensuring shape constraints (e.g., monotonicity), or constructing novel neural network architectures.

Not ideal if your primary goal is simple, off-the-shelf deep learning model training without the need for custom curve-based function approximations or specific shape constraints on your model's outputs.

machine-learning-engineering deep-learning-research feature-engineering model-interpretability neural-networks
Maintenance 10 / 25
Adoption 8 / 25
Maturity 24 / 25
Community 5 / 25

How are scores calculated?

Stars

53

Forks

2

Language

Python

License

Apache-2.0

Last pushed

Feb 17, 2026

Commits (30d)

0

Dependencies

1

Get this data via API

curl "https://pt-edge.onrender.com/api/v1/quality/embeddings/alexshtf/torchcurves"

Open to everyone — 100 requests/day, no key needed. Get a free key for 1,000/day.