tabular-export
Writes the shipped code (src/pipeline.py, src/train.py, src/predict.py) from the winning run and trains it. Usually right after evolve-search or tabular-ensemble. Check: run src/train.py, expects models/model.joblib, metric accuracy (classification) or r2 (regression) with a minimum.
- License 许可
- MIT
- 最近更新
- 2026-10-25
- Stars
- ★ 0
- Forks
- 0
01 前言 Front Matter
- name
- tabular-export
- metadata
- { "beyondml": { "kind": "step", "modality": "tabular" } }
- description
- Writes the shipped code (src/pipeline.py, src/train.py, src/predict.py) from the winning run and trains it. Usually right after evolve-search or tabular-ensemble. Check: run src/train.py, expects models/model.joblib, metric accuracy (classification) or r2 (regression) with a minimum.
- allowed-tools
- read_file list_files write_file run_python export_pipeline finish_step
03 SKILL.md 正文
You turn the winning run into code that works without BeyondML.
- Call export_pipeline(run_id=<the run_id from the upstream summary, or "" for the latest>).
- src/train.py already applies clean(df) and make_features(df) when src/clean.py and src/features.py exist; do not duplicate them.
- Run run_python("src/train.py"). If it fails, read stderr_tail and fix the code (usually a column mismatch between features.py and the exported FEATURES list), then run it again.
- Call finish_step with the reported score, outputs=["src/pipeline.py", "src/train.py", "src/predict.py", "models/model.joblib"].
04 同仓库更多 Skills
tabular-demo
Demo step skill for tabular data.
leakage-demo
Demo knowledge skill about target leakage.
bad-name
Name does not match its folder.
train-test-split
How to split data so the score is honest.
tabular-features
Writes src/features.py with derived features (ratios, logs, interactions) and no target use. Usually right after tabular-eda. Check: run src/features.py, expects data/processed/features.csv.
tabular-ensemble
Combines the search's top models into a stacking ensemble. Only when the top genomes are within about 1% of each other. Check: metric cv_score with a minimum.