ml-model-training

Train ML models by preparing structured data, selecting a framework template, configuring hyperparameters, running the training loop, and evaluating with cross-validation metrics. This Skill provides end-to-end pipeline patterns for classification, regression, and neural networks across all three frameworks.

Prevent overfitting and underfitting through cross-validation, regularization, hyperparameter tuning, and monitoring training vs. validation metrics. This Skill implements explicit handling for both issues with reproducibility controls and convergence checks throughout the training pipeline.

Data preparation and feature engineering involve cleaning, scaling, encoding, and transforming raw structured datasets before training. This Skill includes templates for these preparatory steps integrated into a complete training pipeline with best practices for reproducibility.

Yes. This Skill provides hyperparameter tuning guidance and patterns for all three frameworks within a unified training methodology, enabling model selection and optimization across classification, regression, and neural network tasks on structured data.

Convergence issues stem from learning rate misconfiguration, data preparation errors, or model architecture mismatch. This Skill addresses convergence problems explicitly with best practices for debugging, experiment tracking, and validation metrics to diagnose training failures.