编码

from keras.utils import to_categorical
(x_train,y_train),(x_test,y_test)= fashion_mnist.load_data()
y_train = to_categorical(y_train)

解码

np.argmax(y_train,axis=1)