2 Questions 7 Answers 0 Followers
Questions related from Stefano Di Martino
Hi, this paper https://arxiv.org/pdf/1502.01852.pdf suggests to initialize weights with ReLU and PReLU activations differently. As far as I understand, I initialize the weights of the first layer...
05 May 2017 2,011 4 View
Hi,I trained 32 features with sparse filtering for the MNIST data set. My idea was to use the weights to initialize the first convolutional layer for my deep network which is based on the deep...
04 April 2017 8,952 4 View
