milsonson/anytoany-ml

AnyToAny ML

React + FastAPI numerical AutoML workbench for classification and regression.

Training data uses one sample per line:

1,2,3,4;10,11
2,3,4,5;12,13

The values before ; are features. Values after it are one or more labels. : is also accepted for backward compatibility.

Supported algorithms: MLP, Random Forest, Extra Trees, Gradient Boosting, SVM, KNN, Decision Tree, Logistic Regression, and Linear Regression.

The workbench includes task inference, typed algorithm-specific parameter controls, training progress, holdout charts, feature importance, prediction confidence, and downloadable artifacts.