sklearn的MLPClassifier()的隐藏层和Dense layer of keras/tensorflow一样吗?



理论和实践上,MLPclassifier的隐藏层(参考hidden_layer_sizes)

mlp = MLPClassifier(hidden_layer_sizes=(4, 3, 2, 1),
max_iter = 100, activation = 'relu',
solver = 'adam', verbose = type_spec_from_value,
random_state = 100, learning_rate = 'invscaling',
early_stopping=False
)

与tensorflow/keras的致密层相同

mlp = Sequential()
mlp.add(Dense(4))
mlp.add(Dense(3))
mlp.add(Dense(2))
mlp.add(Dense(3))

?

是的,它们是一样的。在这两种情况下,参数指定神经元的数量。

最新更新