Implementing GB-RBMs for real-valued data

Asked Mar 06 '13 at 12:01

Active Mar 09 '13 at 10:02

Viewed 595 times

I am trying to implement a Deep Belief Network for speech recognition. And hence, need the first layer of RBM to have gaussian visible units.

I used @Andrej Karpathy's matrbm (https://code.google.com/p/matrbm/) and made some changes to convert the Bernouli-Bernouli RBMs to a Gaussian-Bernouli one.

The changed version is available at http://pastebin.com/KTPHnizd

Am I doing anything wrong?

asked Mar 06 '13 at 12:01

Nihar Sarangi

4,845
8
27
32

What is exactly the problem? – Pedrom Mar 11 '13 at 15:02
@Pedrom I am not sure whether the way I am calculating the hidden layer states and visible states (line 109-127) are consistent and correct. – Nihar Sarangi Mar 12 '13 at 06:03
This is more of a request for code review than a problem that someone here can help you solve. – Gus Mar 16 '13 at 00:52

Implementing GB-RBMs for real-valued data

0 Answers0