devbruce / cutmiximagedatagenerator_for_keras Goto Github PK

Will it works with object detection?

Hi, I did the data augmentation for both train and validation data, but I'm error in the below line
history=model.fit_generator(generator=train_generator,validation_data=validation_generator,steps_per_epoch=train_generator.get_steps_per_epoch(),epochs=60,callbacks=[callbacks])

(Error: -> 1122 if validation_data:
1123 val_x, val_y, val_sample_weight = (
1124 data_adapter.unpack_x_y_sample_weight(validation_data))

TypeError: 'NoneType' object cannot be interpreted as an integer)
Any help would be much appreciated!

When I run a simple CNN with CutMixgenerator,

train_gen = CutMixImageDataGenerator(
generator1 = train_generator1,
generator2 = train_generator2,
img_size = image_size,
batch_size = 48,
)
...
model.add(Flatten())
model.add(Dense(2))
model.add(Activation('softmax'))
model.compile(loss='binary_crossentropy', optimizer='adadelta')
model.fit_generator(generator = train_gen, epochs=10, steps_per_epoch=100)

I got an error as follows:

expected activation_5 to have shape (2,) but got array with shape (48,)

Please let me know how to resolve it