conv1d_test.py

import numpy as np
import theano
import theano.tensor as T

X = T.tensor3()
W = theano.shared(np.ones((1,1,1,3)).astype('float32'))
W_shape = (1, 1, 1, 3)

input = X

input_shape = (1, 1, 1, 10)
conved = T.nnet.conv2d(input.dimshuffle(0, 1, 'x', 2), 
                       W,
                       image_shape=input_shape,
                       filter_shape=W_shape,
                       border_mode='valid',
                       ).dimshuffle(0, 1, 3)
input = conved

input_shape = (1, 1, 1, 8)

shift = 1

conved = T.nnet.conv2d(input.dimshuffle(0, 1, 'x', 2), 
                       W,
                       image_shape=input_shape,
                       filter_shape=W_shape,
                       border_mode='full',
                       ).dimshuffle(0, 1, 3)
input = conved[:, :, shift:input.shape[2] + shift]

conved = T.nnet.conv2d(input.dimshuffle(0, 1, 'x', 2), 
                       W,
                       image_shape=input_shape,
                       filter_shape=W_shape,
                       border_mode='full',
                       ).dimshuffle(0, 1, 3)
input = conved[:, :, shift:input.shape[2] + shift]

conved = T.nnet.conv2d(input.dimshuffle(0, 1, 'x', 2), 
                       W,
                       image_shape=input_shape,
                       filter_shape=W_shape,
                       border_mode='full',
                       ).dimshuffle(0, 1, 3)
input = conved[:, :, shift:input.shape[2] + shift]

conved = T.nnet.conv2d(input.dimshuffle(0, 1, 'x', 2), 
                       W,
                       image_shape=input_shape,
                       filter_shape=W_shape,
                       border_mode='full',
                       ).dimshuffle(0, 1, 3)
input = conved[:, :, shift:input.shape[2] + shift]

output = input

f = theano.function([X], output)