CODEWITHSUNDEEP
machine-learningdeep-learningneural-networkcomputer-visionconv-neural-network
Vedant Modi
Vedant Modi

Reputation: 125

ResNet50: Input 0 of layer global_average_pooling2d_2 is incompatible with the layer: expected ndim=4, found ndim=2

Trying to implement ResNet50 on a custom dataset using transfer learning, however get this error:

ValueError: Input 0 of layer global_average_pooling2d_2 is incompatible with the layer: expected ndim=4, found ndim=2. Full shape received: [None, 2048]

Here's my code:

img_height, img_width = (224, 224)
batch_size = 32

train_generator = train_datagen.flow_from_directory(
train_data_dir,
target_size = (img_height, img_width),
batch_size = batch_size,
class_mode = 'categorical',
subset = 'training')


base_model = ResNet50(include_top = False, weights = 'imagenet', pooling='avg')
x = base_model.output
x = GlobalAveragePooling2D()(x)
x = Dense(1024, activation = 'relu')(x)
predictions = Dense(train_generator.num_classes, activation = 'softmax')(x)
model = Model(inputs = base_model.input, outputs = predictions)

for layer in base_model.layers:
    layer.trainable = False
    
model.compile(optimizer = 'adam', loss = 'categorical_crossentropy', metrics = ['accuracy'])

model.fit(train_generator, epochs = 10)

I've set the include_top to False as suggested in some other answers. Where am I going wrong and how do I fix this?

Upvotes: 0

Views: 611

Answers (1)

Adarsh Wase
Adarsh Wase

Reputation: 1890

We don't need GlobalAveragePooling2D in this case, try this code:

base_model = ResNet50(include_top = False, weights = 'imagenet', pooling='avg')
x = base_model.output
x = Dense(1024, activation = 'relu')(x)
predictions = Dense(train_generator.num_classes, activation = 'softmax')(x)
model = Model(inputs = base_model.input, outputs = predictions)

Upvotes: 1

Related Questions