train-model

What problem does it solve?

Implementing and executing machine learning model training loops, including forward/backward passes, gradient updates, and checkpoint management, is a complex and iterative process.

Core Features & Use Cases

• Training Loop Implementation: Guides the setup of data pipelines, model initialization, optimizer configuration, and the core training loop.
• Progress Monitoring: Helps monitor loss values, training time, validation metrics, and checkpoint management.
• Use Case: When developing a new neural network, use this skill to outline a Mojo training loop, including epoch management, loss calculation, and optimization steps, to quickly get your model training.

Quick Start

Use the train-model skill to outline a Mojo struct for a model trainer, including train_epoch functionality.