InvalidArgumentError：断言失败：[Condition x == y did not hold element-wise:]

Question

我正在尝试使用 LSTM 进行分类（进行事件预测）我有 320 个类。数据集由时间序列数据组成，使用以下代码创建和训练：

def create_dataset(dataset,y, look_back=1):
   dataX, dataY = [], []
   for i in range(len(dataset)-look_back):
       a = dataset.iloc[i:(i+look_back)].to_numpy()
       dataX.append(a)
       dataY.append(y.iloc[i + look_back])
   return np.array(dataX), np.array(dataY)

   train_size = int(len(df1) * 0.7)
   test_size = len(df1) - train_size
   train, test = df1.iloc[0:train_size], df1.iloc[train_size:len(df1)]
   print(train.shape, test.shape)
   # reshape into X=t and Y=t+1
   look_back = 10
   #X, Y = create_dataset(dataset, look_back)
   train.msg_code = to_categorical(train.msg_code)
   test.msg_code = to_categorical(test.msg_code)
   trainX, trainY = create_dataset(train,train.msg_code, look_back)
   #print(trainX)
   #print(trainY)
   testX, testY = create_dataset(test,test.msg_code, look_back)
   model = Sequential()
   adam = Adam(lr=0.01)
   #LSTM layers
   chk = ModelCheckpoint('best_model13.pkl', monitor='val_accuracy', save_best_only=True, mode='max', verbose=1)
   model.add(LSTM(128, return_sequences=True, input_shape=(trainX.shape[1],trainX.shape[2])))
   #model.add(LSTM(200, return_sequences=True))
   #model.add(LSTM(200, return_sequences=True))
   #model.add(LSTM(200, return_sequences=True))
   # model.add(LSTM(200, return_sequences=True))
   #Dense layer
   model.add(Dense(320, activation = 'relu'))
   #outputlayer
   model.add(Dense(320,activation='softmax'))
   model.compile(loss='sparse_categorical_crossentropy', optimizer='adam', metrics=['accuracy'])
   #model fitting
   history = model.fit(trainX, trainY,validation_data=(testX,testY), epochs=500, batch_size=32,callbacks=[chk],shuffle=False)
   model = load_model('best_model13.pkl')
   scores = model.evaluate(testX, testY, verbose=0)

我收到以下错误：

InvalidArgumentError:  assertion failed: [Condition x == y did not hold element-wise:] [x (sparse_categorical_crossentropy/SparseSoftmaxCrossEntropyWithLogits/Shape_1:0) = ] [32 1] [y (sparse_categorical_crossentropy/SparseSoftmaxCrossEntropyWithLogits/strided_slice:0) = ] [32 10]
     [[node sparse_categorical_crossentropy/SparseSoftmaxCrossEntropyWithLogits/assert_equal_1/Assert/Assert (defined at <ipython-input-21-2a944dbcd5e3>:220) ]] [Op:__inference_train_function_22113]

Function call stack:
train_function

Answer 1

我在这里找到了答案Tensorflow 2.0 InvalidArgumentError: assertion failed: [Condition x == y did not hold element-wise:]？

我在密集层之前添加了这一行：

model.add(tf.keras.layers.Flatten())