При решении задачи Classification результаты принимают вещественные значения.

Created by: MAGLeb

Решая задачу классификации, результат предсказания не целые числа.

task = 'classification'
composer_params = {
    'max_depth': 2,
    'max_arity': 2,
    'learning_time': 0.1 
}

model = Fedot(ml_task=task,
              composer_params=composer_params)
chain = model.fit(features=np.array(features_train),
                        target=np.array(target_train))
prediction = model.predict(features=np.array(features_valid))

array([9.2810631e-01, 2.1132191e-03, 3.3170132e-03, 9.9691188e-01,
       9.9891078e-01, 9.9755633e-01, 9.9896133e-01, 4.2323591e-04,
       9.9714440e-01, 9.9877244e-01], dtype=float32)

Также хочется обратить внимание на большой LOG при подборе композиций из моделей:

Error in chain assessment during composition: This MLPClassifier estimator requires y to be passed, but the target y is None.. Continue.
Error in chain assessment during composition: Invalid chain configuration: Chain has incorrect models positions. Continue.
Error in chain assessment during composition: Invalid chain configuration: Chain has incorrect models positions. Continue.
Error in chain assessment during composition: This MLPClassifier estimator requires y to be passed, but the target y is None.. Continue.
Error in chain assessment during composition: This MLPClassifier estimator requires y to be passed, but the target y is None.. Continue.
Error in chain assessment during composition: This MLPClassifier estimator requires y to be passed, but the target y is None.. Continue.
Error in chain assessment during composition: This MLPClassifier estimator requires y to be passed, but the target y is None.. Continue.
Error in chain assessment during composition: This MLPClassifier estimator requires y to be passed, but the target y is None.. Continue.
Error in chain assessment during composition: This MLPClassifier estimator requires y to be passed, but the target y is None.. Continue.
Best metric is -0.9916806666666667
Generation num: 0
max_depth: 2, no improvements: 0
Error in chain assessment during composition: This MLPClassifier estimator requires y to be passed, but the target y is None.. Continue.
Error in chain assessment during composition: This MLPClassifier estimator requires y to be passed, but the target y is None.. Continue.
Error in chain assessment during composition: This MLPClassifier estimator requires y to be passed, but the target y is None.. Continue.
spent time: 0.3 min
Best metric is -0.9916806666666667
Composition time: 0.267 min
Algorithm was terminated due to processing time limit
GP composition finished
Fit chain from scratch

Мне кажется некоторые строки не дожны выводиться. Тем более я не ставил параметр verbose. Возможно стоит вынести в отдельный issues