error when run on cpu

[dungdx34]

i have changed the source to run on cpu: use conv2d and pool_2d
but when i run training model, i have a error, please help me:

Traceback (most recent call last):
File "train_bi_char2char.py", line 175, in
main(0, args)
File "train_bi_char2char.py", line 84, in main
prepare_data=prepare_data,
File "/home/dungdx/workspace2/abstract_sentence_convolution/char2char/nmt.py", line 390, in train
cost, not_finite, clipped = f_grad_shared(x, x_mask, y, y_mask)
File "/usr/local/lib/python2.7/dist-packages/theano/compile/function_module.py", line 871, in call
storage_map=getattr(self.fn, 'storage_map', None))
File "/usr/local/lib/python2.7/dist-packages/theano/gof/link.py", line 314, in raise_with_op
reraise(exc_type, exc_value, exc_trace)
File "/usr/local/lib/python2.7/dist-packages/theano/compile/function_module.py", line 859, in call
outputs = self.fn()
File "/usr/local/lib/python2.7/dist-packages/theano/scan_module/scan_op.py", line 951, in rval
r = p(n, [x[0] for x in i], o)
File "/usr/local/lib/python2.7/dist-packages/theano/scan_module/scan_op.py", line 940, in
self, node)
File "theano/scan_module/scan_perform.pyx", line 220, in theano.scan_module.scan_perform.perform (/home/dungdx/.theano/compiledir_Linux-3.13--generic-x86_64-with-Ubuntu-14.04-trusty-x86_64-2.7.6-64/scan_perform/mod.cpp:2442)
ValueError: ('Sequence is shorter then the required number of steps : (n_steps, seq, seq.shape):', 92, array([[ 1., 1., 1., ..., 1., 1., 1.],
[ 1., 1., 1., ..., 1., 1., 1.],
[ 1., 1., 1., ..., 1., 1., 1.],
...,
[ 0., 0., 0., ..., 0., 0., 0.],
[ 0., 0., 0., ..., 0., 0., 0.],
[ 0., 0., 0., ..., 0., 0., 0.]], dtype=float32), (90, 33))
Apply node that caused the error: for{cpu,encoder__layers}(Subtensor{int64}.0, Subtensor{:int64:}.0, Subtensor{:int64:}.0, Subtensor{:int64:}.0, IncSubtensor{Set;:int64:}.0, encoder_U, encoder_Ux)
Toposort index: 805
Inputs types: [TensorType(int64, scalar), TensorType(float32, matrix), TensorType(float32, 3D), TensorType(float32, 3D), TensorType(float32, 3D), TensorType(float32, matrix), TensorType(float32, matrix)]
Inputs shapes: [(), (90, 33), (92, 33, 1024), (92, 33, 512), (93, 33, 512), (512, 1024), (512, 512)]
Inputs strides: [(), (132, 4), (135168, 4096, 4), (67584, 2048, 4), (67584, 2048, 4), (4096, 4), (2048, 4)]
Inputs values: [array(92), 'not shown', 'not shown', 'not shown', 'not shown', 'not shown', 'not shown']
Outputs clients: [[Subtensor{int64::}(for{cpu,encoder__layers}.0, Constant{1}), Elemwise{second,no_inplace}(for{cpu,encoder__layers}.0, DimShuffle{x,x,x}.0)]]

Backtrace when the node is created(use Theano flag traceback.limit=N to make it longer):
File "train_bi_char2char.py", line 175, in
main(0, args)
File "train_bi_char2char.py", line 84, in main
prepare_data=prepare_data,
File "/home/dungdx/workspace2/abstract_sentence_convolution/char2char/nmt.py", line 235, in train
build_model(tparams, model_options)
File "/home/dungdx/workspace2/abstract_sentence_convolution/char2char/char_base.py", line 120, in build_model
proj = get_layer('gru')[1](tparams, hw_out, options, prefix='encoder', mask=x_mask)
File "/home/dungdx/workspace2/abstract_sentence_convolution/char2char/mixer.py", line 614, in gru_layer
strict=True)

HINT: Use the Theano flag 'exception_verbosity=high' for a debugprint and storage map footprint of this apply node.

[jaseleephd]

Hey, the code does not support CPU at the moment. Please refer to #11

[kyunghyuncho]

@jasonleeinf perhaps we should update the code to use the new GPU backend which supports CPU-only convolutional and pooling

https://github.com/Theano/Theano/wiki/Converting-to-the-new-gpu-back-end%28gpuarray%29

[jaseleephd]

@kyunghyuncho Ah cool, didn't know about this. will do this soon!

[dungdx34]

thank you! :)