Then I executed the transformation again, and it gave me another error:
Cell contents reference from a non-cell array object.
'Error in vl_simplenn_tidy (line 37)
layer = net.layers{l} ;
Error in dagnn.DagNN.fromSimpleNN (line 55)
net = vl_simplenn_tidy(net) ;'
So I applied the same correction again.
Then I executed the transform command the third time, and it gave me this error:
'Error using dagnn.DagNN.fromSimpleNN (line 164)
dagnn.Conv is unsupported'
So I went to the line 164, and found this:
'otherwise
error([net.layers{l}.type ' is unsupported']) ;'
And above, on the line 76, there is this:
'case {'conv', 'convt'}'
'case {'conv', 'convt','dagnn.Conv'}'
But now, it pops me this message:
Index exceeds matrix dimensions.
'Error in dagnn.DagNN.fromSimpleNN (line 77)
sz = size(net.layers{l}.weights{1}) ;'
And I have to give up...
So am I doing it right ? How to fix this problem ?
Hi,
I met this problem when I wanted to transform the simple AlexNet NN to DagNN, with the file imagenet-matconvnet-alex:
Matlab command:
alexdag = DagNN.fromSimpleNN(alex)
Error:
`Error in vl_simplenn_move (line 21) switch net.layers{l}.type
Error in dagnn.DagNN.fromSimpleNN (line 54) net = vl_simplenn_move(net, 'cpu') ;'
I searched on the internet, and found this answer: https://fr.mathworks.com/matlabcentral/answers/157686-cell-contents-reference-from-a-non-cell-array-object
So I edited the 21st line of vl_simplenn_move.
Then I executed the transformation again, and it gave me another error:
Error in dagnn.DagNN.fromSimpleNN (line 55) net = vl_simplenn_tidy(net) ;'
So I applied the same correction again.
Then I executed the transform command the third time, and it gave me this error: 'Error using dagnn.DagNN.fromSimpleNN (line 164) dagnn.Conv is unsupported'
So I went to the line 164, and found this: 'otherwise error([net.layers{l}.type ' is unsupported']) ;'
And above, on the line 76, there is this: 'case {'conv', 'convt'}'
'case {'conv', 'convt','dagnn.Conv'}'
But now, it pops me this message:
'Error in dagnn.DagNN.fromSimpleNN (line 77) sz = size(net.layers{l}.weights{1}) ;'
And I have to give up...
So am I doing it right ? How to fix this problem ?
Thanks !