-
Die backpropagation Funktion leakt Speicher. Es werden zum Beispiel immer neue Netze erstellt, die alten aber nicht wieder freigegeben.
Mir ist gerade beim Testen der Speicher vollgelaufen ...
-
It seems like an issue with the used weight update rule (http://ufldl.stanford.edu/wiki/index.php/Backpropagation_Algorithm)
Because once I changed the weight update rule to (http://www.cse.unsw.edu.…
-
The threshold at the LinearActivationStrategy and ThresholdActivationStrategy should be 1 instead of 0. Otherwise the back-propagation of error algorithm will always cancel out the error.
Fokko updated
9 years ago
-
-
Hello,
I'm trying to use cross entropy instead of MSE but I am not sure how to change the perform func. Is it built in?
-
With the current implementation of `LookupTable`, the gradient is computed with respect to all the weights of the lookup table. However we only need to compute the gradient with respect to the weights…
-
I've been using your Boltzmann.jl for my project. I'd like to first appreciate your amazing work. The question I've come across with is
a) How to find the error rate for each iteration.
b) How can I…
-
Has any body madea speed comparation with cuda convnet2 or caffe?I try to reimplement a paper. Using matconvnet is really wonderful.However,I have run 5000 epoch using GPU GTX 550 and 1 million backp…
-
Hi everyone,
I am trying to do activity maximization described in Karen Simonyan's paper (http://arxiv.org/pdf/1312.6034.pdf) using the pre-trained ImageNet model in Caffe. I want to find the input i…
-
Backpropagation is an important technique used for training of neural networks. This task could be started with some basic algorithm and then we'd be able to generalize and refine our design decisions…