def deconvDown(self, input_down, conv_out_up, pooled_out):
# print type(input_down.shape), "type(input_down.shape)"
# print input_down.shape, "input_down.shape"
input_down = theano.shared(input_down, borrow=True)
conv_out_up = theano.shared(conv_out_up, borrow=True)
input_down = T.tanh(input_down)
# conv_out_up = conv_out_up.astype(theano.config.floatX)
# print type(input_down), "type(input_down)", type(pooled_out), type(conv_out_up)
# print input_down.shape, "input_down.shape"
up_pooled = DownsampleFactorMaxGrad(self.poolsize, ignore_border=True,st=None)(conv_out_up.astype(theano.config.floatX), pooled_out, input_down.astype(theano.config.floatX))
output = self.convDown( up_pooled.eval() )
return output
def test_opt_max_to_average(self):
im = theano.tensor.tensor4()
maxout = theano.tensor.tensor4()
grad = theano.tensor.tensor4()
compilation_mode = theano.compile.get_default_mode().including(
'local_average_pool_grad')
for mode in ['max', 'sum', 'average_inc_pad', 'average_exc_pad']:
f = theano.function([im, maxout, grad],
DownsampleFactorMaxGrad(ds=(3, 3),
ignore_border=False,
mode=mode)(im, maxout,
grad),
mode=compilation_mode)
if mode == 'max':
assert any(
isinstance(n.op, MaxPoolGrad)
for n in f.maker.fgraph.toposort())
assert not any(
isinstance(n.op, AveragePoolGrad)
for n in f.maker.fgraph.toposort())
else:
assert not any(
isinstance(n.op, MaxPoolGrad)
for n in f.maker.fgraph.toposort())
assert any(
isinstance(n.op, AveragePoolGrad)
for n in f.maker.fgraph.toposort())
def test_DownsampleFactorMaxGrad(self):
im = theano.tensor.tensor4()
maxout = theano.tensor.tensor4()
grad = theano.tensor.tensor4()
for mode in ['max', 'sum', 'average_inc_pad', 'average_exc_pad']:
f = theano.function([im, maxout, grad],
DownsampleFactorMaxGrad(ds=(3, 3),
ignore_border=False,
mode=mode)(im, maxout,
grad),
on_unused_input='ignore')
if mode == 'max':
assert any(
isinstance(n.op, MaxPoolGrad)
for n in f.maker.fgraph.toposort())
assert not any(
isinstance(n.op, AveragePoolGrad)
for n in f.maker.fgraph.toposort())
else:
assert not any(
isinstance(n.op, MaxPoolGrad)
for n in f.maker.fgraph.toposort())
assert any(
isinstance(n.op, AveragePoolGrad)
for n in f.maker.fgraph.toposort())
def test_infer_shape(self):
image = tensor.dtensor4()
maxout = tensor.dtensor4()
gz = tensor.dtensor4()
rng = numpy.random.RandomState(utt.fetch_seed())
maxpoolshps = ((1, 1), (2, 2), (3, 3), (2, 3), (3, 2))
image_val = rng.rand(4, 6, 7, 9)
out_shapes = [[[4, 6, 7, 9], [4, 6, 7, 9]], [[4, 6, 3, 4],
[4, 6, 4, 5]],
[[4, 6, 2, 3], [4, 6, 3, 3]], [[4, 6, 3, 3],
[4, 6, 4, 3]],
[[4, 6, 2, 4], [4, 6, 3, 5]]]
for i, maxpoolshp in enumerate(maxpoolshps):
for j, ignore_border in enumerate([True, False]):
# checking shapes generated by DownsampleFactorMax
self._compile_and_check([image], [
DownsampleFactorMax(maxpoolshp,
ignore_border=ignore_border)(image)
], [image_val], DownsampleFactorMax)
# checking shapes generated by DownsampleFactorMaxGrad
maxout_val = rng.rand(*out_shapes[i][j])
gz_val = rng.rand(*out_shapes[i][j])
self._compile_and_check([image, maxout, gz], [
DownsampleFactorMaxGrad(maxpoolshp,
ignore_border=ignore_border)(
image, maxout, gz)
], [image_val, maxout_val, gz_val],
DownsampleFactorMaxGrad,
warn=False)
def mp(input, grad):
out = DownsampleFactorMax(maxpoolshp,
ignore_border=ignore_border,
st=stride)(input)
grad_op = DownsampleFactorMaxGrad(
maxpoolshp, ignore_border=ignore_border, st=stride)
return grad_op(input, out, grad)
def __init__(self, bs):
mp = max_pool_2d(bs.output, (2, 2))
self.output = mp
self.output_shape = (bs.output_shape[0], bs.output_shape[1],
bs.output_shape[2] / 2, bs.output_shape[3] / 2)
gradop = DownsampleFactorMaxGrad((2, 2),
st=(2, 2),
padding=(0, 0),
ignore_border=False)
self.upop = lambda v: gradop(bs.output, mp, v)
self.origshape = bs.output_shape
def mp(input, grad):
out = DownsampleFactorMax(
maxpoolsize,
ignore_border=True,
st=stridesize,
padding=paddingsize,
)(input)
grad_op = DownsampleFactorMaxGrad(maxpoolsize,
ignore_border=True,
st=stridesize,
padding=paddingsize)
return grad_op(input, out, grad)
def deconvDown(self, input_down, conv_out_up, pooled_out):
# print type(input_down.shape), "type(input_down.shape)"
# print input_down.shape, "input_down.shape"
input_down = theano.shared(input_down, borrow=True)
conv_out_up = theano.shared(conv_out_up, borrow=True)
input_down = T.maximum(0.0, input_down)
# rect_out = T.maximum(0.0, input_down)
# rect_out = theano.shared(rect_out)
# conv_out_up = conv_out_up.astype(theano.config.floatX)
# print type(input_down), "type(input_down)", type(pooled_out), type(conv_out_up)
# print input_down.shape, "input_down.shape"
# possible to use function from max pooling since it only depends on which value was used in the upward process
up_pooled = DownsampleFactorMaxGrad(self.poolsize, ignore_border=True,st=None)(conv_out_up.astype(theano.config.floatX), pooled_out, input_down.astype(theano.config.floatX))
output = self.convDown( up_pooled.eval() )
return output
def max_pool_2d_same_size(input, patch_size):
output = DownsampleFactorMax(patch_size, True)(input)
outs = DownsampleFactorMaxGrad(patch_size, True)(input, output, output)
return outs
