Deep Learning Experiment Workflow Skill

Overview

Run a staged workflow for deep-learning work where the hard parts are usually investigation quality, experiment definition, and empirical validation rather than large implementation volume. Use this skill for tasks such as model training, fine-tuning, architecture changes, loss-function changes, data-pipeline changes, ablations, benchmark comparisons, and reproducible evaluation work.

This workflow is stage-gated. Do not batch-generate all artifacts by default. Advance only when the current stage gate is satisfied or a classified re-entry path says otherwise.

Skill Layout