Skip to content

curvedinf/k-splanifolds

Repository files navigation

K-Splanifolds

This repo contains the K‑Splanifolds paper plus the code and data used to generate its comparative experiments.

Key files:

  • k-splanifolds.pdf — the paper (reference link).
  • k-splanifolds-2D-to-3D-toy.html — interactive 2D→3D toy visualization.
  • k-splanifolds-3D-to-3D-visualization.html — 3D→3D visualization.
  • k-splanifolds.mp4 — short demo video.
  • setup_venv.sh — creates a local .venv and installs dependencies.

Key experiment code:

  • experiments/common/splanifold.py — core splanifold map and utilities used by the experiments.
  • experiments/runtime_baselines/run.py — runtime baselines (Table 2 + Figure 1 data).
  • experiments/curve_fitting/run.py — curve fitting under a fixed optimization budget (Table 3 + Figure 2 data).
  • experiments/entropy_bits/run.py — entropy-coded bits per parameter in regression (Table 4 + Figure 10 data).

Setup & Running Experiments

Setup:

./setup_venv.sh

Running:

source .venv/bin/activate
python experiments/runtime_baselines/run.py

About

Linear-scaling compute & memory spline manifold for regression and interpolation

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

 
 
 

Contributors