Using learned weights for network initilization

[ssakhavi]

Hi,

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

[skaae]

I think you manually have to move the weights from the RBM object to the NN object.
The RBM notation is:

% NOTATION:
% data : all data given as [n_samples x #vis]
% W : vis - hid weights [ #hid x #vis ]
% U : label - hid weights [ #hid x #n_classes ]
% b : bias of visible layer [ #vis x 1]
% c : bias of hidden layer [ #hid x 1]
% d : bias of label layer [ #n_classes x 1]

On 02 Nov 2014, at 08:25, Siavash Sakhavi notifications@github.com wrote:

Hi,

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

—
Reply to this email directly or view it on GitHub.

[ssakhavi]

Thanks

On Sun, Nov 2, 2014 at 4:30 PM, Søren Kaae Sønderby <
notifications@github.com> wrote:

I think you manually have to move the weights from the RBM object to the
NN object.
The RBM notation is:

% NOTATION:
% data : all data given as [n_samples x #vis]
% W : vis - hid weights [ #hid x #vis ]
% U : label - hid weights [ #hid x #n_classes ]
% b : bias of visible layer [ #vis x 1]
% c : bias of hidden layer [ #hid x 1]
% d : bias of label layer [ #n_classes x 1]

On 02 Nov 2014, at 08:25, Siavash Sakhavi notifications@github.com
wrote:

Hi,

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

Reply to this email directly or view it on GitHub.

Reply to this email directly or view it on GitHub
#2 (comment).