def createModelBase(self, weights='imagenet'):
model_name_for_call = self.decideModelName()
modelCall = getattr(self, model_name_for_call)
model, img_width, img_height = modelCall(weights)
for layer in model.layers:
print(layer.name)
layer.trainable = True
if weights == 'imagenet':
model.layers.pop()
pretrained_inputs = model.inputs
model = Flatten()(model.output)
model = Dense(512, activation='relu')(model)
model = Dropout(0.5)(model)
predictions = Dense(self.NUMBER_OF_CLASSES,
activation='softmax')(model)
else:
print(model.summary())
pretrained_inputs = model.input
predictions = Dense(self.NUMBER_OF_CLASSES,
activation='softmax',
name='dense_1')(model.layers[-2].output)
print(model.summary())
model_final = Model(inputs=pretrained_inputs, outputs=predictions)
return model_final, img_height, img_width
# FC layers
model = Flatten()(model)
#model = Dense(1024, kernel_regularizer=l2(0.01), bias_regularizer=l2(0.01))(model)
model = Dense(1024)(model)
#model = Dropout(0.2)(model)
#model = Dense(64, kernel_regularizer=l2(0.01), bias_regularizer=l2(0.01))(model)
model = Dense(64)(model)
#model = Dropout(0.2)(model)
output = Dense(num_classes, activation='softmax')(model)
model = Model(inputs=[input_image], outputs=[output])
print(model.summary())
# Compile Model
model.compile(optimizer='Adam',
loss='categorical_crossentropy',
metrics=['accuracy'])
'''
train model
'''
batch_size = 256
num_epochs = 20
# Train Model
history = model.fit(trainX, trainY, batch_size=batch_size,
epochs=num_epochs) #, callbacks=[checkpoint])
model = Dense(dense_neurons / 2, activation='tanh')(model)
# Output Layer
output = Dense(10, activation="softmax")(model)
model = Model(input_layer, output)
# Compiling model
optimizer = keras.optimizers.SGD(lr=learning_rate, momentum=momentum)
model.compile(
loss="sparse_categorical_crossentropy",
optimizer=optimizer,
metrics=["accuracy"]
)
model.summary()
# Train the model
history = model.fit(
train_data,
epochs=13,
validation_data = test_data
)
"""# Saving and Recreating the trained model"""
## Save the whole model
model.save('./trained_CNN/imagewoof/my_model_imagewoof.h5')
## Recreate whole model
new_model=keras.models.load_model('./trained_CNN/imagewoof/my_model_imagewoof.h5')
