You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Hi majianjia.
With your kind support of ZeroPadding layer, I deployed my caffe model successfully.
Now I am trying to deploy my model on memory constraint chip, and I found ZeroPadding layer consume much memory.
I think ZeroPadding and Cropping can be replaced as Single Buffer layer, not sure though.
I expect a quick response as before.
Thanks again for your great work.
The text was updated successfully, but these errors were encountered:
Hi @Alan-Turing-Ko,
Thanks for your suggestion, the current method is to keep them simple and efficient.
I think it is possible to make them single buffer layers.
Cropping is straight forward, but zero padding need to be done in reverse order.
I am planning to do it in the following days.
If you cant wait, you can also make your version and pull request are always welcome.
Jianjia
Hi majianjia.
With your kind support of ZeroPadding layer, I deployed my caffe model successfully.
Now I am trying to deploy my model on memory constraint chip, and I found ZeroPadding layer consume much memory.
#3 Conv2D - ReLU - ( 64, 64, 8) 819k (17,424,32,768, 100) 1 1 1 - - - - -
#4 ZeroPad - - ( 65, 65, 8) (32,768,33,800, 0) 1 - 1 - - - - -
I think ZeroPadding and Cropping can be replaced as Single Buffer layer, not sure though.
I expect a quick response as before.
Thanks again for your great work.
The text was updated successfully, but these errors were encountered: