81 lines
2.6 KiB
Python
81 lines
2.6 KiB
Python
from sacred import Experiment
|
|
from sacred.observers import MongoObserver
|
|
import os
|
|
import mlflow
|
|
import mlflow.keras
|
|
from mlflow.models.signature import infer_signature
|
|
from mlflow.models import Model
|
|
|
|
os.environ["SACRED_NO_GIT"] = "1"
|
|
|
|
ex = Experiment('s487187-training', interactive=True, save_git_info=False)
|
|
ex.observers.append(MongoObserver(url='mongodb://admin:IUM_2021@172.17.0.1:27017', db_name='sacred'))
|
|
|
|
@ex.config
|
|
def my_config():
|
|
data_file = 'data.csv'
|
|
model_file = 'model'
|
|
epochs = 10
|
|
batch_size = 32
|
|
test_size = 0.2
|
|
random_state = 42
|
|
|
|
|
|
@ex.capture
|
|
def train_model(data_file, model_file, epochs, batch_size, test_size, random_state):
|
|
import pandas as pd
|
|
from sklearn.model_selection import train_test_split
|
|
from sklearn.preprocessing import MinMaxScaler
|
|
import tensorflow as tf
|
|
from imblearn.over_sampling import SMOTE
|
|
|
|
with mlflow.start_run():
|
|
|
|
smote = SMOTE(random_state=random_state)
|
|
data = pd.read_csv(data_file, sep=';')
|
|
|
|
data = pd.get_dummies(data, columns=['Sex', 'Medal'])
|
|
data = data.drop(columns=['Name', 'Team', 'NOC', 'Games', 'Year', 'Season', 'City', 'Sport', 'Event'])
|
|
|
|
scaler = MinMaxScaler()
|
|
data = pd.DataFrame(scaler.fit_transform(data), columns=data.columns)
|
|
|
|
X = data.filter(regex='Sex|Age')
|
|
y = data.filter(regex='Medal')
|
|
y = pd.get_dummies(y)
|
|
|
|
X = X.fillna(0)
|
|
y = y.fillna(0)
|
|
|
|
y = y.values
|
|
|
|
X_resampled, y_resampled = smote.fit_resample(X, y)
|
|
X_train, X_test, y_train, y_test = train_test_split(X_resampled, y_resampled, test_size=test_size, random_state=random_state)
|
|
|
|
model = tf.keras.models.Sequential()
|
|
model.add(tf.keras.layers.Dense(64, input_dim=X_train.shape[1], activation='relu'))
|
|
model.add(tf.keras.layers.Dense(32, activation='relu'))
|
|
model.add(tf.keras.layers.Dense(y.shape[1], activation='softmax'))
|
|
|
|
model.compile(loss='categorical_crossentropy', optimizer='adam', metrics=['accuracy'])
|
|
|
|
model.fit(X_train, y_train, epochs=epochs, batch_size=batch_size)
|
|
loss, accuracy = model.evaluate(X_test, y_test)
|
|
|
|
mlflow.log_metric("loss", loss)
|
|
mlflow.log_metric("accuracy", accuracy)
|
|
|
|
signature = infer_signature(X_train, model.predict(X_train))
|
|
input_example = Model.log_input_example(X_train.iloc[0])
|
|
|
|
mlflow.keras.log_model(model, model_file, signature=signature, input_example=input_example)
|
|
|
|
return accuracy
|
|
|
|
@ex.main
|
|
def run_experiment():
|
|
accuracy = train_model()
|
|
ex.log_scalar('accuracy', accuracy)
|
|
ex.add_artifact('model.h5')
|
|
|
|
ex.run() |