skaae / rbm_toolbox_lua Goto Github PK

Hi,

I wanted to ask how I can use the output rbm object for initializing a network in nn

Hi,

I was wondering if this library had any support for GPU ?

Thanks

In the paper "Learning Algorithms for the Classification Restricted Boltzmann Machine", a uniform distribution in the interval [-m**-0.5, m**-0.5] is used for the initialisation of the weight matrices (rbm.W and rbm.U).

However, in the initcrbm the following code is used:

M = math.max(m,n);
interval_max = math.pow(M,-0.5);
interval_min = -interval_max;
weights = torch.rand(m,n):mul( interval_min + (interval_max-interval_min) )

which initialises the weights using the interval [0, m**-0.5].

Should the calculation of the weights not be?

weights = torch.rand(m,n):mul(interval_max-interval_min):add(interval_min)

Hi,
I have recently downloded this package and while wanting to see some of the examples, the mnist.createdatasets is retuning a r value.
I don't know whether it is an issue with the dataset link or with the method itself.

Hi,
After having problems with my own rbm implementation I was very happy to notice your code, and wanted to try out the examples.
Unfortunately the http://data.neuflow.org/data/mnist-th7.tgz link is broken and you dont seem to use the same format as the dataset in https://s3.amazonaws.com/torch7/data/mnist.t7.tgz