Use early_stopping or not?

[guanjen375]

I have noticed the XGB training is as follows:

for x in tqdm(range(100)):
    x_train, x_test, y_train, y_test = train_test_split(data, margin, test_size=.1)

    train = xgb.DMatrix(x_train, label=y_train)
    test = xgb.DMatrix(x_test, label=y_test)

    param = {
        'max_depth': 2,
        'eta': 0.01,
        'objective': 'multi:softprob',
        'num_class': 2
    }
    epochs = 500

    model = xgb.train(param, train, epochs)
    predictions = model.predict(test)
    y = []

    for z in predictions:
        y.append(np.argmax(z))

    acc = round(accuracy_score(y_test, y)*100, 1)

I have also noticed some people do not use the forloop to train the models:

x_train, x_test, y_train, y_test = train_test_split(data, margin, test_size=.1)

train = xgb.DMatrix(x_train, label=y_train)
test = xgb.DMatrix(x_test, label=y_test)

param = {
    'max_depth': 2,
    'eta': 0.01,
    'objective': 'multi:softprob',
    'num_class': 2
}
epochs = 500
eval_set = [(x_test,y_test)]

model = xgb.train(param, train, epochs, early_stopping_rounds=10, eval_set = eval_set )

predictions = model.predict(test)
y = []

for z in predictions:
    y.append(np.argmax(z))

acc = round(accuracy_score(y_test, y)*100, 1)

Which one do you prefer? Or there is another better way to train the model?