Lemm Ras questions

Lemm Ras answers

• (matlab) MLP with relu and softmax not working with mini-batch SGD and produces similar predictions on MNIST dataset
• ReLU Function In A Recurrent Neural Network. Weights Become Infinity or Zero
• What's the difference between Softmax and SoftmaxWithLoss layer in caffe?
• Deep neural network skip connection implemented as summation vs concatenation? [closed]