src.train_logistic_classifier

train_logistic_classifier.py

This script trains a logistic regression classifier on latent representations obtained from an encoder model. It loads latent vectors and their corresponding binary labels, splits them into training and testing subsets, trains a LogisticRegression model using scikit-learn, evaluates its performance (accuracy, classification report, and optionally confusion matrix), and saves the trained model to disk in .pkl format.

Usage:

python train_logistic_classifier.py –latents X_latents.npy –labels y_labels.npy –output model.pkl [–plot]

Example

python train_logistic_classifier.py –latents data/X_latents.npy –labels data/y_labels.npy –output output/logistic_model.pkl –test_size 0.25 –max_iter 500 –seed 123 –plot

Functions

train_logistic_classifier(X_path, y_path, ...)

Train and evaluate a logistic regression classifier on latent representations.