Open wanghonghui1998 opened 3 years ago
Thank you for your reminder, I will modify this file, but it seems to have little effect?
That's OK, I just wonder if the auxiliary head is used or not during search.
I also noticed that the 'cutout' augment is set to be true by default in train_search.py. Hence, Both autoaugment and cutout are used in searching step?
Hi!Thanks for your code. I noticed that the 'auxiliary' augment is set to be true by default in train_search.py. That seems to be different from GOLD-NAS paper.