Skip to main content
Shameek ConyersMarketing Data ScienceResumeContact

Models

Small model implementations for studying fit and prediction APIs.

Supervised models over SicLib tensors

This module contains small supervised learning models built on SicLib tensors. The current API covers k-nearest neighbors and a linear model. I use these implementations to test the tensor layer and study fit and prediction interfaces, not for production workloads.

Documentation

class KNearestNeighbors:
    def __init__(self) -> None: ...
    def fit_model(self, x_vals: _pysiclib.linalg.Tensor, y_vals: _pysiclib.linalg.Tensor, num_neighbors: int, metric: int = ...) -> None: ...
    def predict(self, arg0: _pysiclib.linalg.Tensor) -> _pysiclib.linalg.Tensor: ...

class LinearModel:
    def __init__(self) -> None: ...
    def fit_model(self, x_vals: _pysiclib.linalg.Tensor, y_vals: _pysiclib.linalg.Tensor, fit_procedure: int = ...) -> None: ...
    def predict(self, arg0: _pysiclib.linalg.Tensor) -> _pysiclib.linalg.Tensor: ...