upura / atma-comp5

atmaCup #5

• atmaCup #5のソースコード。最終順位はpublic 16位->private 27位。
• 取り組みについては、ブログやapproach.mdをご覧ください。
• 自作コンペ用ライブラリ「Ayniy」を利用しています。

Documentation | GitHub | Slide (Japanese)

# Import packages
import yaml
from sklearn.model_selection import StratifiedKFold
from ayniy.preprocessing.runner import Tabular
from ayniy.model.runner import Runner

# Load configs
f = open('configs/fe000.yml', 'r+')
fe_configs = yaml.load(f)
g = open('configs/run000.yml', 'r+')
run_configs = yaml.load(g)

# Difine CV strategy as you like
cv = StratifiedKFold(n_splits=5, shuffle=True, random_state=7)

# Feature engineering
tabular = Tabular(fe_configs, cv)
tabular.create()

# Modeling
runner = Runner(run_configs, cv)
runner.run_train_cv()
runner.run_predict_cv()
runner.submission()

Environment

docker-compose build
docker-compose up

MLflow

cd experiments
mlflow ui

Test

docker-compose build
docker-compose up -d
docker exec -it ayniy-test bash

pytest tests/ --cov=. --cov-report=html

Docs

docker-compose build
docker-compose up -d
docker exec -it ayniy-test bash
cd docs
make html

cd docs/build/html
git a .
git c "update"
git push origin master

https://github.com/upura/ayniy-docs