search

layumi / 2016_super_resolution

ICCV2015 Image Super-Resolution Using Deep Convolutional Networks
MIT License
85 stars 43 forks source link

Error when testing the code. #2

Open faris060 opened 7 years ago

faris060 commented 7 years ago

Hi, i'm interested in your work on super resolution. I encounter a problem when running testSRnet_result.m. the error is regarding vl-nnconv saying that an input is not a numeric array (or GPU support not compiled.

The rest of the error are as follow:

Error using vl_nnconv An input is not a numeric array (or GPU support not compiled).

`Error` in dagnn.Conv/forward (line 11)
      outputs{1} = vl_nnconv(...
`Error` in dagnn.Layer/forwardAdvanced (line 85)
      outputs = obj.forward(inputs, {net.params(par).value})
  ;
Error` in dagnn.DagNN/eval (line 91)
  obj.layers(l).block.forwardAdvanced(obj.layers(l)) ;
`Error` in testSRnet_result (line 27)
net.eval({'input',gpuArray(input_big)});

Do i need to change any parameter during compilation?

layumi commented 7 years ago

@faris060 It seems that your matlab doses't recognise the 'gpuArray'. I strongly recommend you refer to http://www.vlfeat.org/matconvnet/install/#compiling Or you may try to run my code on CPU. Use vlnn_compile() to compile. Then change all the 'gpuArray' format data to 'single' format. But it may take a while.