def build_pool(self, layer, previous_layer, last_output, **kwargs):
entity = Pool(kwargs['shape'])
if previous_layer['name'] == 'Convolution':
layer['output_shape'] = entity.output_shape(
previous_layer['output_shape']
)
else:
layer['output_shape'] = (
previous_layer['output_shape'] / kwargs['shape']
)
### Logging ###
self.logger.info("pool output shape")
self.logger.info(layer['output_shape'])
return (layer, entity)
conv0 = Convolution.withFilters(filter_shape=(nkernels[0], 1, 8, 8),
image_shape=(
batch_size,
1,
) + t3_images.shape[-2:],
filters=filters_multi)
fm0 = conv0.get_output(x)
next_filter = conv0.pipe_filter_shape(20, 4, 4)
next_img = conv0.pipe_image_shape
print("Next_filter conv0")
print(next_filter)
print(next_img)
print("Pool0")
pool0 = Pool((2, 2))
pool_out0 = pool0.get_output(fm0)
next_filter = pool0.pipe_filter_shape((nkernels[0], 1, 8, 8), 20, 4, 4)
next_img = pool0.pipe_image_shape((
batch_size,
1,
) + t3_images.shape[-2:], (nkernels[0], 1, 8, 8))
print("Next_filter pool0")
print(next_filter)
print(next_img)
print("####")
print((nkernels[1], nkernels[0], 4, 4))
print((batch_size, nkernels[0], 12, 12))
conv1 = Convolution.withoutFilters(filter_shape=(nkernels[1], nkernels[0], 4,
4),
image_shape=(batch_size, nkernels[0], 12,
conv0 = Convolution.withFilters(
filter_shape=(nkernels[0], 1, 8, 8),
image_shape=(batch_size, 1,) + t3_images.shape[-2:],
filters=filters_multi
)
fm0 = conv0.get_output(x)
next_filter = conv0.pipe_filter_shape(20, 4, 4)
next_img = conv0.pipe_image_shape
print("Next_filter conv0")
print(next_filter)
print(next_img)
print("Pool0")
pool0 = Pool((2,2))
pool_out0 = pool0.get_output(fm0)
next_filter = pool0.pipe_filter_shape((nkernels[0], 1, 8, 8), 20, 4, 4)
next_img = pool0.pipe_image_shape(
(batch_size, 1,) + t3_images.shape[-2:],
(nkernels[0], 1, 8, 8)
)
print("Next_filter pool0")
print(next_filter)
print(next_img)
print("####")
print((nkernels[1],nkernels[0],4,4))
print((batch_size, nkernels[0], 12, 12))
conv1 = Convolution.withoutFilters(
filter_shape=(nkernels[1],nkernels[0],4,4),
image_shape=(batch_size, nkernels[0], 12, 12)
