Once deciding on a few constant labels, can one freeze the current weights of the models, to ensure a quicker inference time?
Once deciding on a few constant labels, can one freeze the current weights of the models, to ensure a quicker inference time?