vat_tf icon indicating copy to clipboard operation
vat_tf copied to clipboard

Why not update statistics in BatchNorm layer?

Open SpinachR opened this issue 7 years ago • 0 comments

Hi,

I found that in your 'build_training_graph' function in train_semisup.py. The 'update_batch_stats' set to False, in this case, the avg_mean and avg_var would always be 0 and 1 during the EVAL.

I am not sure about it or Is there any trick in this setting.

SpinachR avatar Feb 03 '18 04:02 SpinachR