Wonder about lrn parameters.

[sicnarf1a]

Hi.
I am curious about the 'local response normalization' layer parameters you set.
The paper case is a little bit different from what you set except 'beta'.

What does it mean to be different from the actual setting?

[matteo-dunnhofer]

Hello @enterai,

for the hyperparameters I just followed the working pre-trained implementation here. Did not have the chance to test other parameter setups. If you have the opportunity than let me know the results 👍🏻.