Skip to content

README.md

Latest commit

 

History

History
50 lines (38 loc) · 1.77 KB

README.md

File metadata and controls

50 lines (38 loc) · 1.77 KB
Raw

Python examples

Runnable Python hosts for the hypercube-esn API. They use only the public package surface — no CMake, no C++ example binaries.

Script What it shows
basic_prediction.py Scalar next-step regression (fit → R² / NRMSE)
classification.py Binary labels + classification head (accuracy)

How to run

Examples live in the git tree. They are not installed by the PyPI wheel (pip install hypercube-esn alone does not place these scripts on disk).

From the repository root (after a Release wheel or local build):

pip install hypercube-esn
# or, from this tree:  pip install ./python
python python/examples/basic_prediction.py
python python/examples/classification.py

From the python/ directory after an editable/local install:

cd python
pip install . --no-build-isolation   # needs a C++ toolchain if no wheel cache
python examples/basic_prediction.py

What these are not

  • Not the frozen C++ storefronts (NARMA best-5, MemoryCapacity grids, Lorenz VPT / GS surveys). Those live under examples/ and report the paper / release numbers.
  • Not pytest. CI smoke is tests/test_basic.py; these scripts are for humans to read and run.
  • Not hard tasks. Sine next-step and sign-of-sine classification are easy onboarding signals so the API is obvious; do not cite their metrics as storefront results.

Going further

Want… See…
Full Python API docs/Python_SDK.md
C++ BasicPrediction / NARMA / Lorenz examples/README.md
Package install / PyPI README.md