google-colaboratorytf.kerasvalueerrorprecision-recall

ValueError: Shapes (None, 1) and (None, 5) are incompatible in keras

model = Sequential()

model.add(Conv2D(128, (3, 3), activation='relu', input_shape=(64, 64, 3), padding='same'))
model.add(MaxPooling2D(pool_size=(2, 2)))

model.add(Conv2D(64, (3, 3), activation='relu', padding='same'))
model.add(MaxPooling2D(pool_size=(2, 2)))

model.add(Conv2D(32, (3, 3), activation='relu', padding='same'))
model.add(MaxPooling2D(pool_size=(2, 2)))

model.add(Flatten())
model.add(Dense(128, activation = 'relu'))
model.add(Dropout(0.5))
model.add(Dense(5, activation = 'softmax'))


model.compile(optimizer='adam', loss='sparse_categorical_crossentropy', metrics=[tf.keras.metrics.Recall()])

This code works fine for metrics=['accuracy']), but it shows ValueError: Shapes (None, 1) and (None, 5) are incompatible for metrics=[tf.keras.metrics.Recall()])

Please help me. Thanks in advance.

Solution

  • Recall makes sense only for binary classification. Your final layer has 5 nodes, which essentially means you have 5 classes. You should change recall to another metric. Documentation should help you choose an appropriate metric for your model. Categorical accuracy should be good enough to get started.