Skip to content

no sandbox.cuda #109

@yuval-harpaz

Description

@yuval-harpaz

Hi
following the Guided Backpropagation recipe/example I had error for:
if theano.sandbox.cuda.cuda_enabled because, I guess, I use newer theano / lasagne versions with CuDNN etc. I have no cuda under theano.sandbox. two solutions:

  1. skip the if and use the else option: maybe_to_gpu = lambda x: x
  2. use gpu with gpuarray.dnn
    maybe_to_gpu = theano.gpuarray.dnn.as_gpuarray_variable x = maybe_to_gpu(x,None)
    and then a few lines below (as above, add None as second input argument):
    outp = maybe_to_gpu(self.nonlinearity(inp),None)
    Thanks for the examples

Metadata

Metadata

Assignees

No one assigned

    Labels

    No labels
    No labels

    Type

    No type

    Projects

    No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions