The error is still now in LightCapsNet. Please help me to fix the bug or update the conv_2d. file in LightCapsNet.

Please help me to fix the following bug in LightCapsNet:
Undefined function or variable
'PADDING_MODE'.

Error in conv_layer_2d (line 60)
if PADDING_MODE==1

Error in net_bp (line 39)
[res(layer).dzdx,
res(layer).dzdw,res(layer).dzdb,opts]
= conv_layer_2d(
res(layer).x,net.layers{layer}.weights{1},net.layers{layer}.weights{2},net.layers{layer}.stride,net.layers{layer}.pad,res(layer+1).dzdx,opts
);

Error in select_learning_rate (line
59)
[ net,res,opts ] =
net_bp( net,res,opts
);

Error in selective_sgd (line 13)
[lr_best]=select_learning_rate(net,opts);

Error in train_net (line 26)
[ net,opts ] = selective_sgd(
net,opts );

Error in TrainingScript (line 46)
[net,opts]=train_net(net,opts);

Error in Main_Template (line 128)
TrainingScript();

Error in Main_MNIST_Conv_Capsules
(line 30)
Main_Template(); %call training
template

Array dimensions must match for binary array op.

Error in conv_layer_2d (line 115)
y=y+bias_p;

Error in net_ff (line 39)
[res(layer+1).x,,,opts] = conv_layer_2d(
res(layer).x,net.layers{layer}.weights{1},net.layers{layer}.weights{2},net.layers{layer}.stride,net.layers{layer}.pad,[],opts
);

Error in select_learning_rate (line 52)
[ net,res,opts ] = net_ff( net,res,opts );

Error in selective_sgd (line 13)
[lr_best]=select_learning_rate(net,opts);

Error in train_net (line 26)
[ net,opts ] = selective_sgd( net,opts );

Error in TrainingScript (line 46)
[net,opts]=train_net(net,opts);

Error in Main_Template (line 128)
TrainingScript();

Error in Main_MNIST_Conv_Capsules (line 30)
Main_Template(); %call training template

Error in RunAll (line 10)
Main_MNIST_Conv_Capsules()

my problem is in testing the network, I have tried to use " predict " command but it does not predict.

predict (net,x)
Error using predict (line 126)
No valid system or dataset was specified.

this is the message.

Hi,thank you very much for your code,I ran your code but it gave me the error:

Undefined variable "nnet" or class "nnet.internal.cnn.layer.Convolution2D"
conv_layer_2d (line 36)
opts.layer{opts.current_layer}.conv2d_nntb=nnet.internal.cnn.layer.Convolution2D('conv2d_nntb', [k1,k2], k3,k4, [stride(1),stride(2)], pad(1:2:end));

I can't find the solution about it and I hope I can get your generous help.
thanks a lot!

Hi,

I tried to run the RunAll and I got the following error:

>> RunAll
Training a non-convolutional CapsNet.

use_gpu =

     1


opts = 

    use_nntoolbox: 1

Undefined variable "nnet" or class "nnet.internal.cnn.layer.Convolution2D".

Error in linear_layer (line 17)
            opts.layer{opts.current_layer}.fc_nntb=nnet.internal.cnn.layer.Convolution2D('conv2d_nntb', [1,1], in ,out, [1,1], [0,0]);

Error in net_ff (line 56)
                [res(layer+1).x,~,~,opts] = linear_layer( res(layer).x,net.layers{layer}.weights{1},net.layers{layer}.weights{2},[], opts );

Error in select_learning_rate (line 52)
                [ net,res,opts ] = net_ff( net,res,opts );

Error in selective_sgd (line 13)
        [lr_best]=select_learning_rate(net,opts);

Error in train_net (line 26)
        [ net,opts ] = selective_sgd( net,opts );

Error in TrainingScript (line 46)
    [net,opts]=train_net(net,opts);

Error in Main_Template (line 128)
    TrainingScript();

Error in Main_MNIST_MLP_Capsules (line 30)
Main_Template(); %call training template

Error in RunAll (line 4)
Main_MNIST_MLP_Capsules();

Does anyone know why I'm getting this error?
Best regards.

EDIT

This issue was already answered here

Hi there,

Awesome repository. I've been trying to implement an MLP based caps net, where the primary caps analyzes subsets of features in groups, feeding them into MLPs per capsule then pushed to downstream capsules.

For some reason, I am getting terrible results. Do you know what I might be doing wrong?

What are areas that make MLP capsules different from CNN capsules, what should I be aware of? How do I optimize to get good results. I think dynamic routing is essential for establishment of hierarchy, but in the non-image context, I'm unsure of how the affine transformation is helpful.

My implementation is in pytorch.