FewShotContinualLearning
FewShotContinualLearning copied to clipboard
about data.py
Thanks for your great job! I have some doubt about the data sampling within a class_change_interval. Why not remove the picked selected_classes*(self.num_samples_per_support_class+self.num_samples_per_target_class) after one iteration of support set and target set generation, in order to avoid the duplicate samples among different support sets within a class_change_interval?
Hi @liujianzhao6328057 can you please be more specific? Please report the name of the files and the lines you are referring to.