检查目标时出错：预期 activation_6 有 3 个维度，但得到了形状为 (70612, 1) 的数组

Question

I Keep getting the error: Error when checking target: expected activation_6 to have 3 dimensions, but got array with shape (70612, 1). What could be the issue?

see below code:

# CNN Model
model = Sequential()
model.add(Embedding(256, 8))

model.add(Conv1D(32, 3, activation="relu"))
model.add(MaxPooling1D(pool_size = (2)))
model.add(Conv1D(32, 3, activation="relu"))
model.add(MaxPooling1D(pool_size =(2)))


model.add(Dense(64)
model.add(Dropout(0.5))
model.add(Dense(2))
model.add(Activation("softmax"))

model.compile(loss="categorical_crossentropy", optimizer ="adam", metrics = ["accuracy"])

**Model Summary:**

Model: "sequential_37"
_________________________________________________________________
Layer (type)                 Output Shape              Param #   
=================================================================
embedding_37 (Embedding)     (None, None, 8)           2048      
_________________________________________________________________
conv1d_47 (Conv1D)           (None, None, 32)          800       
_________________________________________________________________
max_pooling1d_43 (MaxPooling (None, None, 32)          0         
_________________________________________________________________
conv1d_48 (Conv1D)           (None, None, 32)          3104      
_________________________________________________________________
max_pooling1d_44 (MaxPooling (None, None, 32)          0         
_________________________________________________________________
dense_34 (Dense)             (None, None, 64)          2112      
_________________________________________________________________
dropout_18 (Dropout)         (None, None, 64)          0         
_________________________________________________________________
dense_35 (Dense)             (None, None, 2)           130       
_________________________________________________________________
activation_12 (Activation)   (None, None, 2)           0         
=================================================================

Answer 1

Your last layer has 3 dimension, it should have been 2 (batch, n_class) .

activation_12 (Activation) (None, None, 2) 0

You need to add a Flatten() layer before

model.add(Dense(64)

Another issue is you have your labels with shape (batch, 1) .

You need to use sigmoid instead of softmax (assuming binary classification).

Also, change the shape of x and y.

x = x.reshape((-1, 533, 1)) y = y.reshape(-1,1)

model = Sequential()
model.add(Embedding(256, 8, input_length = 533))

model.add(Conv1D(32, 3, activation="relu"))
model.add(MaxPooling1D(pool_size = (2)))
model.add(Conv1D(32, 3, activation="relu"))
model.add(MaxPooling1D(pool_size =(2)))

model.add(Flatten())
model.add(Dense(64)
model.add(Dropout(0.5))
model.add(Dense(1)) # 1
model.add(Activation("sigmoid"))

model.compile(loss="binary_crossentropy", optimizer ="adam", metrics = ["accuracy"])