train_label = tf.keras.backend.one_hot(train_label,3) train_label = tf.one_hot(train_label,3)
выдает следующую ошибку в tenorflow 2.0.0
InternalError: Could not find valid device for node.
Узел: {{node OneHot}}
train_label должно быть типом данных INT.например:
train_label
INT
train_label = [1, 2 ,3] train_label = tf.one_hot(train_label,3) // work train_label = [1., 2., 3.] train_label = tf.one_hot(train_label,3) // InternalError: Could not find valid device for node.