analyze-results

To analyze ML experiment results, this skill aggregates metrics from common output directories, normalizes the data, and computes statistical summaries including mean, standard deviation, and delta versus baseline across multiple seeds and parameter sweeps.

Consolidating machine learning model runs involves gathering outputs from directories like figures/ or results/, normalizing them into a structured dataset, and generating comparative trends to highlight significant improvements or regressions.

Generating a statistical report for ML hyperparameter sweep outputs requires aggregating metrics, computing statistical summaries to flag anomalies, and producing an insights section with observations and recommended next steps suitable for documentation.

Yes, you can use this to find baseline improvements and regressions by computing delta versus baseline metrics, building comparison tables, and automatically surfacing concise observations about significant trends in your machine learning experiments.

When analyzing machine learning experiment outputs, the statistics computed include mean, standard deviation, and delta versus baseline, alongside anomaly flagging across various runs and seeds to derive actionable insights for research decisions.

No specific frameworks are required to aggregate ML model metrics; the approach identifies experiment outputs saved in common directories like figures/ or results/, normalizes the diverse formats into a structured dataset, and automatically computes comparative trends.