Its great code implementation. Thank you. I have one question that in model.fit why training labels are zeros

[gashishoo7]

I have one question that why labels are zeros in model.fit training funtion.
model.fit(
x=[train_set_x1, train_set_x2],
y=np.zeros(len(train_set_x1)),
batch_size=batch_size,
epochs=epoch_num,
shuffle=True,
validation_data=([valid_set_x1, valid_set_x2], np.zeros(len(valid_set_x1))),
callbacks=[checkpointer],
)

[sumit7692]

Because DCCA is a self-supervised model and does not require labels, but due to constraints of model.fit, y=np.zeros(len(train_set_x1)) is used. If you go through the objective function, training labels are not used. Those are just dummy values.