pacnet
pacnet copied to clipboard
Reason for No ReLU
Is there any intuition behind having no ReLU layers after the last conv layers and all of the pac_t layers?
Sorry for the delay! There are ReLU layers after pac_t layers. See https://github.com/NVlabs/pacnet/blob/master/task_jointUpsampling/models.py#L248
The output of the last conv is the output of the network. Depending on the task, it usually doesn't make sense to clip values at zero.