3 tagged experiment-management, measured the same way as everything else here.
Skill Claude CodeCodex
Use when operating a labrat lab with Codex: checking health, choosing the next phase prompt, supervising runtime cycles, auditing candidates, synthesizing recent evaluations, or writing checkpoint notes.
Skill Claude CodeCodex
Migrate an existing project that uses pytorchtemplate to the latest version. Use this skill when the user says: "update template", "migrate", "upgrade", "apply latest changes", "sync with template", "update pytorchtemplate", or when the user's project is missing new features (preflight, hpo-report, data field…
Skill Claude CodeCodex
Guide for training PyTorch models using this template's config-driven pipeline. Use this skill whenever the user wants to: train a model, create experiment configs (run.yaml / opt.yaml / best.yaml), run hyperparameter optimization (HPO) with Optuna, extract best parameters from an HPO study, set up a new experiment…