Tune with scaled batch size

[WangDaHa]

Hi Dr. Jia,

I noticed that in tune_fgvc/tune_vtab.py, the learning rate is scaled as 【lr = lr / 256 * cfg.DATA.BATCH_SIZE】for choosing the best learning rate while in train.py, this operation is not used (lr is kept as [5, 10, 50, etc]). I wonder the reason for doing this. Are the reported results based on the unscaled learning rate? (lr is given in your fgvc excel file)

Best,

[Menglin Jia]

hi, thanks for the question. we use tune*.py for all of our experiments, so we set the scaled lr in the tune_vtab.py file. So the reported results are all based on the scaled learning rate.

train.py does not set the scaling operation since it is only used as a main gate way for training. If you want to use this one directly, please remember to do the scaling manually.

[WangDaHa]

Thanks for your information! I will close the question.