我有多个输出
out = [Dense(19, name='one', activation='softmax')(out), Dense(19, name='two', activation='softmax')(out), Dense(19, name='three', activation='softmax')(out), Dense(19, name='four', activation='softmax')(out)]model.fit(reshape_train_X, y_onehot, batch_size=400, epochs=100, verbose=2, validation_split=0.2, callbacks=callbacks_list)这是我的y_onehot格式：
[array([[1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0], [0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0], [0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0], [0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0]], dtype=uint8), array([[1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0], [0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0], [0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0], [0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0]],dtype=uint8),.....]我得到了这个错误信息
ValueError: Error when checking model target: the list of Numpy arrays that you are passing to your model is not the size the model expected. Expected to see 4 array(s), but instead got the following list of 5000 arrays: [array([[1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0], [0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0], [0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0], ...我不知道为什么当y_onehot中有四个列表时会出现这个错误。
len(y_onehot): 5000
print(“y_onehot”, y_onehot[0])
[[1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0] [0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0] [0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0] [0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0]]print(“y_onehot”, len(y_onehot[0]))
y_onehot 4我尝试了这个。但仍然没有效果。
感谢您的帮助。
回答：
这是一个示例。请注意您的y。您必须在fit中分别传递每个输出
inp = Input((50))x = Dense(32)(inp)x1 = Dense(19, name='one', activation='softmax')(x)x2 = Dense(19, name='two', activation='softmax')(x)x3 = Dense(19, name='three', activation='softmax')(x)x4 = Dense(19, name='four', activation='softmax')(x)model = Model(inp, [x1,x2,x3,x4])model.compile('adam', 'categorical_crossentropy')X = np.random.uniform(0,1, (5000,50))y1 = np.random.randint(0,2, (5000,19))y2 = np.random.randint(0,2, (5000,19))y3 = np.random.randint(0,2, (5000,19))y4 = np.random.randint(0,2, (5000,19))model.fit(X, [y1,y2,y3,y4], epochs=10)