def test_leaky_relu(self):
"""Test creation of a LeakyReLU activation"""
activation_name = 'LeakyReLU'
args = {}
activation = activation_factory.create(activation_name, **args)
self.assertEqual(activation._get_name(), activation_name)
def test_softmax(self):
"""Test creation of a Softmax activation"""
activation_name = 'Softmax'
args = {}
activation = activation_factory.create(activation_name, **args)
self.assertEqual(activation._get_name(), activation_name)
x = torch.empty(10, 2)
y = activation(x)
assert_array_equal(y, torch.softmax(x, -1))
def test_sigmoid(self):
"""Test creation of a Sigmoid activation"""
activation_name = 'Sigmoid'
args = {}
activation = activation_factory.create(activation_name, **args)
self.assertEqual(activation._get_name(), activation_name)
x = torch.empty(10)
y = activation(x)
assert_array_equal(y, torch.sigmoid(x))
def test_relu(self):
"""Test creation of a ReLU activation"""
activation_name = 'ReLU'
args = {}
activation = activation_factory.create(activation_name, **args)
self.assertEqual(activation._get_name(), activation_name)
x = torch.ones(10) * -1
y = activation(x)
self.assertEqual(len(torch.nonzero(y, as_tuple=False)), 0)
def __init__(self, block, layers, in_channels=3,
zero_init_residual=False, groups=1, width_per_group=64,
replace_stride_with_dilation=None, norm_layer=None,
activation={'name': 'ReLU', 'params': {'inplace': True}}):
super(ResNet, self).__init__()
if norm_layer is None:
norm_layer = nn.BatchNorm2d
self._norm_layer = norm_layer
self.inplanes = 64
self.dilation = 1
if replace_stride_with_dilation is None:
# each element in the tuple indicates if we should replace
# the 2x2 stride with a dilated convolution instead
replace_stride_with_dilation = [False, False, False]
if len(replace_stride_with_dilation) != 3:
raise ValueError("replace_stride_with_dilation should be None "
"or a 3-element tuple, got {}".format(
replace_stride_with_dilation))
self.groups = groups
self.base_width = width_per_group
self.conv1 = nn.Conv2d(in_channels, self.inplanes, kernel_size=7,
stride=2, padding=3, bias=False)
self.bn1 = norm_layer(self.inplanes)
self.activation = activation_factory.create(
activation['name'], **activation['params'])
self.maxpool = nn.MaxPool2d(kernel_size=3, stride=2, padding=1)
self.layer1 = self._make_layer(block, 64, layers[0])
self.layer2 = self._make_layer(block, 128, layers[1], stride=2,
dilate=replace_stride_with_dilation[0])
self.layer3 = self._make_layer(block, 256, layers[2], stride=2,
dilate=replace_stride_with_dilation[1])
self.layer4 = self._make_layer(block, 512, layers[3], stride=2,
dilate=replace_stride_with_dilation[2])
for m in self.modules():
if isinstance(m, nn.Conv2d):
nn.init.kaiming_normal_(m.weight, mode='fan_out', nonlinearity='relu')
elif isinstance(m, (nn.BatchNorm2d, nn.GroupNorm)):
nn.init.constant_(m.weight, 1)
nn.init.constant_(m.bias, 0)
# Zero-initialize the last BN in each residual branch,
# so that the residual branch starts with zeros, and each residual
# block behaves like an identity. This improves the model by 0.2~0.3%
# according to https://arxiv.org/abs/1706.02677
if zero_init_residual:
for m in self.modules():
if isinstance(m, Bottleneck):
nn.init.constant_(m.bn3.weight, 0)
elif isinstance(m, BasicBlock):
nn.init.constant_(m.bn2.weight, 0)