def build_inception_module(name, input_layer, nfilters):
# nfilters: (pool_proj, 1x1, 3x3_reduce, 3x3, 5x5_reduce, 5x5)
net = dict()
net['pool'] = PoolLayerDNN(input_layer, pool_size=3, stride=1, pad=1)
net['pool_proj'] = ConvLayer(net['pool'],
nfilters[0],
1,
flip_filters=False)
net['1x1'] = ConvLayer(input_layer, nfilters[1], 1, flip_filters=False)
net['3x3_reduce'] = ConvLayer(input_layer,
nfilters[2],
1,
flip_filters=False)
net['3x3'] = ConvLayer(net['3x3_reduce'],
nfilters[3],
3,
pad=1,
flip_filters=False)
net['5x5_reduce'] = ConvLayer(input_layer,
nfilters[4],
1,
flip_filters=False)
net['5x5'] = ConvLayer(net['5x5_reduce'],
nfilters[5],
5,
pad=2,
flip_filters=False)
net['output'] = ConcatLayer(
[net['1x1'], net['3x3'], net['5x5'], net['pool_proj']])
return {'{}/{}'.format(name, k): v for k, v in net.items()}
def build_inception_module(name, prefix, input_layer, nfilters):
# nfilters: (pool_proj, 1x1, 3x3_reduce, 3x3, 5x5_reduce, 5x5)
subnet = {}
subnet['pool'] = PoolLayerDNN(input_layer,
pool_size=3,
stride=1,
pad=1,
name=prefix + name + '/pool')
subnet['pool_proj'] = ConvLayer(subnet['pool'],
nfilters[0],
1,
flip_filters=False,
name=prefix + name + '/pool_proj')
subnet['1x1'] = ConvLayer(input_layer,
nfilters[1],
1,
flip_filters=False,
name=prefix + name + '/1x1')
subnet['3x3_reduce'] = ConvLayer(input_layer,
nfilters[2],
1,
flip_filters=False,
name=prefix + name + '/3x3_reduce')
subnet['3x3'] = ConvLayer(subnet['3x3_reduce'],
nfilters[3],
3,
pad=1,
flip_filters=False,
name=prefix + name + '/3x3')
subnet['5x5_reduce'] = ConvLayer(input_layer,
nfilters[4],
1,
flip_filters=False,
name=prefix + name + '/5x5_reduce')
subnet['5x5'] = ConvLayer(subnet['5x5_reduce'],
nfilters[5],
5,
pad=2,
flip_filters=False,
name=prefix + name + '/5x5')
subnet['output'] = ConcatLayer([
subnet['1x1'],
subnet['3x3'],
subnet['5x5'],
subnet['pool_proj'],
],
name=prefix + name + '/output')
return {prefix + '{}/{}'.format(name, k): v for k, v in subnet.items()}
def build_model(self, input_var, forward, dropout):
net = dict()
net['input'] = InputLayer((None, 3, None, None), input_var=input_var)
net['conv1/7x7_s2'] = ConvLayer(net['input'],
64,
7,
stride=2,
pad=3,
flip_filters=False)
net['pool1/3x3_s2'] = PoolLayer(net['conv1/7x7_s2'],
pool_size=3,
stride=2,
ignore_border=False)
net['pool1/norm1'] = LRNLayer(net['pool1/3x3_s2'], alpha=0.00002, k=1)
net['conv2/3x3_reduce'] = ConvLayer(net['pool1/norm1'],
64,
1,
flip_filters=False)
net['conv2/3x3'] = ConvLayer(net['conv2/3x3_reduce'],
192,
3,
pad=1,
flip_filters=False)
net['conv2/norm2'] = LRNLayer(net['conv2/3x3'], alpha=0.00002, k=1)
net['pool2/3x3_s2'] = PoolLayerDNN(net['conv2/norm2'],
pool_size=3,
stride=2)
net.update(
self.build_inception_module('inception_3a', net['pool2/3x3_s2'],
[32, 64, 96, 128, 16, 32]))
net.update(
self.build_inception_module('inception_3b',
net['inception_3a/output'],
[64, 128, 128, 192, 32, 96]))
net['pool3/3x3_s2'] = PoolLayerDNN(net['inception_3b/output'],
pool_size=3,
stride=2)
net.update(
self.build_inception_module('inception_4a', net['pool3/3x3_s2'],
[64, 192, 96, 208, 16, 48]))
net.update(
self.build_inception_module('inception_4b',
net['inception_4a/output'],
[64, 160, 112, 224, 24, 64]))
net.update(
self.build_inception_module('inception_4c',
net['inception_4b/output'],
[64, 128, 128, 256, 24, 64]))
net.update(
self.build_inception_module('inception_4d',
net['inception_4c/output'],
[64, 112, 144, 288, 32, 64]))
net.update(
self.build_inception_module('inception_4e',
net['inception_4d/output'],
[128, 256, 160, 320, 32, 128]))
net['pool4/3x3_s2'] = PoolLayerDNN(net['inception_4e/output'],
pool_size=3,
stride=2)
net.update(
self.build_inception_module('inception_5a', net['pool4/3x3_s2'],
[128, 256, 160, 320, 32, 128]))
net.update(
self.build_inception_module('inception_5b',
net['inception_5a/output'],
[128, 384, 192, 384, 48, 128]))
net['pool5/7x7_s1'] = GlobalPoolLayer(net['inception_5b/output'])
if forward:
#net['fc6'] = DenseLayer(net['pool5/7x7_s1'], num_units=1000)
net['prob'] = DenseLayer(net['pool5/7x7_s1'],
num_units=4,
nonlinearity=softmax)
else:
net['dropout1'] = DropoutLayer(net['pool5/7x7_s1'], p=dropout)
#net['fc6'] = DenseLayer(net['dropout1'], num_units=1000)
#net['dropout2'] = DropoutLayer(net['fc6'], p=dropout)
net['prob'] = DenseLayer(net['dropout1'],
num_units=4,
nonlinearity=softmax)
return net