autoresearch

What problem does it solve?

This Skill automates the process of iterating on machine learning model training, allowing for continuous improvement without manual intervention.

Core Features & Use Cases

• Iterative Model Improvement: Automatically modifies training code, runs experiments, evaluates metrics, and keeps only beneficial changes.
• Cross-Platform Support: Works across CUDA, Apple Silicon MPS, and CPU environments.
• Use Case: Let an AI agent autonomously discover better model architectures or hyperparameters for your deep learning project overnight, reporting back with the best-performing configuration.

Quick Start

Use the autoresearch skill to autonomously run an ML research loop on train.py with a 5-minute time budget per experiment, optimizing for validation accuracy.