ML Model Trainer Skill for Reproducible Classification
Train on labeled synthetic records with a fixed split, keep model selection inside training, and save the model and evaluation report.
中文Install in Claude Code
Use Python, pandas, scikit-learn, and synthetic parcel width, weight, and size labels. No external model service is needed.
- Get the complete `plugins/ai-ml/ml-model-trainer/skills/training-machine-learning-models` directory from the pinned GitHub commit.
- Copy it into the target project's `.claude/skills/`. Check `.claude/skills/training-machine-learning-models/SKILL.md`, start Claude Code at the project root, and call `/training-machine-learning-models`.
Example task
Create 200 parcels marked small or medium. Use 80% for training and cross-validation. Evaluate once on the remaining 20%, then save the model and holdout report.What to check
The report identifies the split, seed, macro F1, and class support. A saved artifact can replay the local example prediction.
Original SKILL.md on GitHub · Updated
Frequently asked question
Should I repeatedly check the test split when tuning parameters?
Choose parameters with validation inside the training set. Reserve the held-out set for final evaluation.