IO Error:Requires a GPU to work.

[vishwajit s.]

When I try to do:

from lasagne.layers.corrmm import Conv2DMMLayer as ConvLayer

it returns 

Requires a GPU to work:( as corrmm.py doesn't have a mode for CPUs)

Although my theano (config.device) recognizes that I have a CPU.

[Jan Schlüter]

When I try to do:

from lasagne.layers.corrmm import Conv2DMMLayer as ConvLayer

it returns 

Requires a GPU to work:( as corrmm.py doesn't have a mode for CPUs)

Easiest solution: Do "from lasagne.layers import Conv2DLayer as ConvLayer" instead. You usually don't need the Conv2DMMLayer, Theano will pick the fastest available implementation by itself.

 

Although my theano (config.device) recognizes that I have a CPU.

If you meant "GPU", and Theano doesn't use it, see the installation guide for your platform: https://github.com/Lasagne/Lasagne/wiki/From-Zero-to-Lasagne

Good luck!

Jan

[vishwajit s.]

Hey Jan, 
That worked, thanks but now I get a memory error when I try to build the VGG 19 convnet as I import the 7th fully connected layer with 4096 units. 
I'll post the details in a few hours.

Thanks in advance!

[Jan Schlüter]

That worked, thanks but now I get a memory error when I try to build the VGG 19 convnet as I import the 7th fully connected layer with 4096 units. 
I'll post the details in a few hours.

Probably your GPU is too small -- how much memory does it have?

[leepe...@gmail.com]

Hi, can you please send me the vgg19.pkl file?