Python train_evaluate Exemples

Exemple #1

Fichier : train.py Projet : neptune-ai/blog-hyperparameter_optimization

import pandas as pd

from utils import train_evaluate

N_ROWS = 10000
TRAIN_PATH = '/mnt/ml-team/minerva/open-solutions/santander/data/train.csv'
MODEL_PARAMS = {
    'boosting': 'gbdt',
    'objective': 'binary',
    'metric': 'auc',
    'num_threads': 12,
    'learning_rate': 0.3,
}

data = pd.read_csv(TRAIN_PATH, nrows=N_ROWS)

X = data.drop(['ID_code', 'target'], axis=1)
y = data['target']

score = train_evaluate(X, y, MODEL_PARAMS)
print('Validation AUC: {}'.format(score))

Exemple #2

Fichier : search_hyperopt_tpe.py Projet : neptune-ai/blog-hyperparameter_optimization

def objective(params):
    all_params = {**params, **STATIC_PARAMS}
    return -1.0 * train_evaluate(X, y, all_params)