in easycv/models/backbones/bninception.py [0:0]
def __init__(self, num_classes=0):
super(BNInception, self).__init__()
inplace = True
self.conv1_7x7_s2 = nn.Conv2d(
3, 64, kernel_size=(7, 7), stride=(2, 2), padding=(3, 3))
self.conv1_7x7_s2_bn = nn.BatchNorm2d(64, affine=True)
self.conv1_relu_7x7 = nn.ReLU(inplace)
self.pool1_3x3_s2 = nn.MaxPool2d((3, 3),
stride=(2, 2),
dilation=(1, 1),
ceil_mode=True)
self.conv2_3x3_reduce = nn.Conv2d(
64, 64, kernel_size=(1, 1), stride=(1, 1))
self.conv2_3x3_reduce_bn = nn.BatchNorm2d(64, affine=True)
self.conv2_relu_3x3_reduce = nn.ReLU(inplace)
self.conv2_3x3 = nn.Conv2d(
64, 192, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
self.conv2_3x3_bn = nn.BatchNorm2d(192, affine=True)
self.conv2_relu_3x3 = nn.ReLU(inplace)
self.pool2_3x3_s2 = nn.MaxPool2d((3, 3),
stride=(2, 2),
dilation=(1, 1),
ceil_mode=True)
self.inception_3a_1x1 = nn.Conv2d(
192, 64, kernel_size=(1, 1), stride=(1, 1))
self.inception_3a_1x1_bn = nn.BatchNorm2d(64, affine=True)
self.inception_3a_relu_1x1 = nn.ReLU(inplace)
self.inception_3a_3x3_reduce = nn.Conv2d(
192, 64, kernel_size=(1, 1), stride=(1, 1))
self.inception_3a_3x3_reduce_bn = nn.BatchNorm2d(64, affine=True)
self.inception_3a_relu_3x3_reduce = nn.ReLU(inplace)
self.inception_3a_3x3 = nn.Conv2d(
64, 64, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
self.inception_3a_3x3_bn = nn.BatchNorm2d(64, affine=True)
self.inception_3a_relu_3x3 = nn.ReLU(inplace)
self.inception_3a_double_3x3_reduce = nn.Conv2d(
192, 64, kernel_size=(1, 1), stride=(1, 1))
self.inception_3a_double_3x3_reduce_bn = nn.BatchNorm2d(
64, affine=True)
self.inception_3a_relu_double_3x3_reduce = nn.ReLU(inplace)
self.inception_3a_double_3x3_1 = nn.Conv2d(
64, 96, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
self.inception_3a_double_3x3_1_bn = nn.BatchNorm2d(96, affine=True)
self.inception_3a_relu_double_3x3_1 = nn.ReLU(inplace)
self.inception_3a_double_3x3_2 = nn.Conv2d(
96, 96, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
self.inception_3a_double_3x3_2_bn = nn.BatchNorm2d(96, affine=True)
self.inception_3a_relu_double_3x3_2 = nn.ReLU(inplace)
self.inception_3a_pool = nn.AvgPool2d(
3, stride=1, padding=1, ceil_mode=True, count_include_pad=True)
self.inception_3a_pool_proj = nn.Conv2d(
192, 32, kernel_size=(1, 1), stride=(1, 1))
self.inception_3a_pool_proj_bn = nn.BatchNorm2d(32, affine=True)
self.inception_3a_relu_pool_proj = nn.ReLU(inplace)
self.inception_3b_1x1 = nn.Conv2d(
256, 64, kernel_size=(1, 1), stride=(1, 1))
self.inception_3b_1x1_bn = nn.BatchNorm2d(64, affine=True)
self.inception_3b_relu_1x1 = nn.ReLU(inplace)
self.inception_3b_3x3_reduce = nn.Conv2d(
256, 64, kernel_size=(1, 1), stride=(1, 1))
self.inception_3b_3x3_reduce_bn = nn.BatchNorm2d(64, affine=True)
self.inception_3b_relu_3x3_reduce = nn.ReLU(inplace)
self.inception_3b_3x3 = nn.Conv2d(
64, 96, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
self.inception_3b_3x3_bn = nn.BatchNorm2d(96, affine=True)
self.inception_3b_relu_3x3 = nn.ReLU(inplace)
self.inception_3b_double_3x3_reduce = nn.Conv2d(
256, 64, kernel_size=(1, 1), stride=(1, 1))
self.inception_3b_double_3x3_reduce_bn = nn.BatchNorm2d(
64, affine=True)
self.inception_3b_relu_double_3x3_reduce = nn.ReLU(inplace)
self.inception_3b_double_3x3_1 = nn.Conv2d(
64, 96, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
self.inception_3b_double_3x3_1_bn = nn.BatchNorm2d(96, affine=True)
self.inception_3b_relu_double_3x3_1 = nn.ReLU(inplace)
self.inception_3b_double_3x3_2 = nn.Conv2d(
96, 96, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
self.inception_3b_double_3x3_2_bn = nn.BatchNorm2d(96, affine=True)
self.inception_3b_relu_double_3x3_2 = nn.ReLU(inplace)
self.inception_3b_pool = nn.AvgPool2d(
3, stride=1, padding=1, ceil_mode=True, count_include_pad=True)
self.inception_3b_pool_proj = nn.Conv2d(
256, 64, kernel_size=(1, 1), stride=(1, 1))
self.inception_3b_pool_proj_bn = nn.BatchNorm2d(64, affine=True)
self.inception_3b_relu_pool_proj = nn.ReLU(inplace)
self.inception_3c_3x3_reduce = nn.Conv2d(
320, 128, kernel_size=(1, 1), stride=(1, 1))
self.inception_3c_3x3_reduce_bn = nn.BatchNorm2d(128, affine=True)
self.inception_3c_relu_3x3_reduce = nn.ReLU(inplace)
self.inception_3c_3x3 = nn.Conv2d(
128, 160, kernel_size=(3, 3), stride=(2, 2), padding=(1, 1))
self.inception_3c_3x3_bn = nn.BatchNorm2d(160, affine=True)
self.inception_3c_relu_3x3 = nn.ReLU(inplace)
self.inception_3c_double_3x3_reduce = nn.Conv2d(
320, 64, kernel_size=(1, 1), stride=(1, 1))
self.inception_3c_double_3x3_reduce_bn = nn.BatchNorm2d(
64, affine=True)
self.inception_3c_relu_double_3x3_reduce = nn.ReLU(inplace)
self.inception_3c_double_3x3_1 = nn.Conv2d(
64, 96, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
self.inception_3c_double_3x3_1_bn = nn.BatchNorm2d(96, affine=True)
self.inception_3c_relu_double_3x3_1 = nn.ReLU(inplace)
self.inception_3c_double_3x3_2 = nn.Conv2d(
96, 96, kernel_size=(3, 3), stride=(2, 2), padding=(1, 1))
self.inception_3c_double_3x3_2_bn = nn.BatchNorm2d(96, affine=True)
self.inception_3c_relu_double_3x3_2 = nn.ReLU(inplace)
self.inception_3c_pool = nn.MaxPool2d((3, 3),
stride=(2, 2),
dilation=(1, 1),
ceil_mode=True)
self.inception_4a_1x1 = nn.Conv2d(
576, 224, kernel_size=(1, 1), stride=(1, 1))
self.inception_4a_1x1_bn = nn.BatchNorm2d(224, affine=True)
self.inception_4a_relu_1x1 = nn.ReLU(inplace)
self.inception_4a_3x3_reduce = nn.Conv2d(
576, 64, kernel_size=(1, 1), stride=(1, 1))
self.inception_4a_3x3_reduce_bn = nn.BatchNorm2d(64, affine=True)
self.inception_4a_relu_3x3_reduce = nn.ReLU(inplace)
self.inception_4a_3x3 = nn.Conv2d(
64, 96, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
self.inception_4a_3x3_bn = nn.BatchNorm2d(96, affine=True)
self.inception_4a_relu_3x3 = nn.ReLU(inplace)
self.inception_4a_double_3x3_reduce = nn.Conv2d(
576, 96, kernel_size=(1, 1), stride=(1, 1))
self.inception_4a_double_3x3_reduce_bn = nn.BatchNorm2d(
96, affine=True)
self.inception_4a_relu_double_3x3_reduce = nn.ReLU(inplace)
self.inception_4a_double_3x3_1 = nn.Conv2d(
96, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
self.inception_4a_double_3x3_1_bn = nn.BatchNorm2d(128, affine=True)
self.inception_4a_relu_double_3x3_1 = nn.ReLU(inplace)
self.inception_4a_double_3x3_2 = nn.Conv2d(
128, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
self.inception_4a_double_3x3_2_bn = nn.BatchNorm2d(128, affine=True)
self.inception_4a_relu_double_3x3_2 = nn.ReLU(inplace)
self.inception_4a_pool = nn.AvgPool2d(
3, stride=1, padding=1, ceil_mode=True, count_include_pad=True)
self.inception_4a_pool_proj = nn.Conv2d(
576, 128, kernel_size=(1, 1), stride=(1, 1))
self.inception_4a_pool_proj_bn = nn.BatchNorm2d(128, affine=True)
self.inception_4a_relu_pool_proj = nn.ReLU(inplace)
self.inception_4b_1x1 = nn.Conv2d(
576, 192, kernel_size=(1, 1), stride=(1, 1))
self.inception_4b_1x1_bn = nn.BatchNorm2d(192, affine=True)
self.inception_4b_relu_1x1 = nn.ReLU(inplace)
self.inception_4b_3x3_reduce = nn.Conv2d(
576, 96, kernel_size=(1, 1), stride=(1, 1))
self.inception_4b_3x3_reduce_bn = nn.BatchNorm2d(96, affine=True)
self.inception_4b_relu_3x3_reduce = nn.ReLU(inplace)
self.inception_4b_3x3 = nn.Conv2d(
96, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
self.inception_4b_3x3_bn = nn.BatchNorm2d(128, affine=True)
self.inception_4b_relu_3x3 = nn.ReLU(inplace)
self.inception_4b_double_3x3_reduce = nn.Conv2d(
576, 96, kernel_size=(1, 1), stride=(1, 1))
self.inception_4b_double_3x3_reduce_bn = nn.BatchNorm2d(
96, affine=True)
self.inception_4b_relu_double_3x3_reduce = nn.ReLU(inplace)
self.inception_4b_double_3x3_1 = nn.Conv2d(
96, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
self.inception_4b_double_3x3_1_bn = nn.BatchNorm2d(128, affine=True)
self.inception_4b_relu_double_3x3_1 = nn.ReLU(inplace)
self.inception_4b_double_3x3_2 = nn.Conv2d(
128, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
self.inception_4b_double_3x3_2_bn = nn.BatchNorm2d(128, affine=True)
self.inception_4b_relu_double_3x3_2 = nn.ReLU(inplace)
self.inception_4b_pool = nn.AvgPool2d(
3, stride=1, padding=1, ceil_mode=True, count_include_pad=True)
self.inception_4b_pool_proj = nn.Conv2d(
576, 128, kernel_size=(1, 1), stride=(1, 1))
self.inception_4b_pool_proj_bn = nn.BatchNorm2d(128, affine=True)
self.inception_4b_relu_pool_proj = nn.ReLU(inplace)
self.inception_4c_1x1 = nn.Conv2d(
576, 160, kernel_size=(1, 1), stride=(1, 1))
self.inception_4c_1x1_bn = nn.BatchNorm2d(160, affine=True)
self.inception_4c_relu_1x1 = nn.ReLU(inplace)
self.inception_4c_3x3_reduce = nn.Conv2d(
576, 128, kernel_size=(1, 1), stride=(1, 1))
self.inception_4c_3x3_reduce_bn = nn.BatchNorm2d(128, affine=True)
self.inception_4c_relu_3x3_reduce = nn.ReLU(inplace)
self.inception_4c_3x3 = nn.Conv2d(
128, 160, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
self.inception_4c_3x3_bn = nn.BatchNorm2d(160, affine=True)
self.inception_4c_relu_3x3 = nn.ReLU(inplace)
self.inception_4c_double_3x3_reduce = nn.Conv2d(
576, 128, kernel_size=(1, 1), stride=(1, 1))
self.inception_4c_double_3x3_reduce_bn = nn.BatchNorm2d(
128, affine=True)
self.inception_4c_relu_double_3x3_reduce = nn.ReLU(inplace)
self.inception_4c_double_3x3_1 = nn.Conv2d(
128, 160, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
self.inception_4c_double_3x3_1_bn = nn.BatchNorm2d(160, affine=True)
self.inception_4c_relu_double_3x3_1 = nn.ReLU(inplace)
self.inception_4c_double_3x3_2 = nn.Conv2d(
160, 160, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
self.inception_4c_double_3x3_2_bn = nn.BatchNorm2d(160, affine=True)
self.inception_4c_relu_double_3x3_2 = nn.ReLU(inplace)
self.inception_4c_pool = nn.AvgPool2d(
3, stride=1, padding=1, ceil_mode=True, count_include_pad=True)
self.inception_4c_pool_proj = nn.Conv2d(
576, 128, kernel_size=(1, 1), stride=(1, 1))
self.inception_4c_pool_proj_bn = nn.BatchNorm2d(128, affine=True)
self.inception_4c_relu_pool_proj = nn.ReLU(inplace)
self.inception_4d_1x1 = nn.Conv2d(
608, 96, kernel_size=(1, 1), stride=(1, 1))
self.inception_4d_1x1_bn = nn.BatchNorm2d(96, affine=True)
self.inception_4d_relu_1x1 = nn.ReLU(inplace)
self.inception_4d_3x3_reduce = nn.Conv2d(
608, 128, kernel_size=(1, 1), stride=(1, 1))
self.inception_4d_3x3_reduce_bn = nn.BatchNorm2d(128, affine=True)
self.inception_4d_relu_3x3_reduce = nn.ReLU(inplace)
self.inception_4d_3x3 = nn.Conv2d(
128, 192, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
self.inception_4d_3x3_bn = nn.BatchNorm2d(192, affine=True)
self.inception_4d_relu_3x3 = nn.ReLU(inplace)
self.inception_4d_double_3x3_reduce = nn.Conv2d(
608, 160, kernel_size=(1, 1), stride=(1, 1))
self.inception_4d_double_3x3_reduce_bn = nn.BatchNorm2d(
160, affine=True)
self.inception_4d_relu_double_3x3_reduce = nn.ReLU(inplace)
self.inception_4d_double_3x3_1 = nn.Conv2d(
160, 192, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
self.inception_4d_double_3x3_1_bn = nn.BatchNorm2d(192, affine=True)
self.inception_4d_relu_double_3x3_1 = nn.ReLU(inplace)
self.inception_4d_double_3x3_2 = nn.Conv2d(
192, 192, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
self.inception_4d_double_3x3_2_bn = nn.BatchNorm2d(192, affine=True)
self.inception_4d_relu_double_3x3_2 = nn.ReLU(inplace)
self.inception_4d_pool = nn.AvgPool2d(
3, stride=1, padding=1, ceil_mode=True, count_include_pad=True)
self.inception_4d_pool_proj = nn.Conv2d(
608, 128, kernel_size=(1, 1), stride=(1, 1))
self.inception_4d_pool_proj_bn = nn.BatchNorm2d(128, affine=True)
self.inception_4d_relu_pool_proj = nn.ReLU(inplace)
self.inception_4e_3x3_reduce = nn.Conv2d(
608, 128, kernel_size=(1, 1), stride=(1, 1))
self.inception_4e_3x3_reduce_bn = nn.BatchNorm2d(128, affine=True)
self.inception_4e_relu_3x3_reduce = nn.ReLU(inplace)
self.inception_4e_3x3 = nn.Conv2d(
128, 192, kernel_size=(3, 3), stride=(2, 2), padding=(1, 1))
self.inception_4e_3x3_bn = nn.BatchNorm2d(192, affine=True)
self.inception_4e_relu_3x3 = nn.ReLU(inplace)
self.inception_4e_double_3x3_reduce = nn.Conv2d(
608, 192, kernel_size=(1, 1), stride=(1, 1))
self.inception_4e_double_3x3_reduce_bn = nn.BatchNorm2d(
192, affine=True)
self.inception_4e_relu_double_3x3_reduce = nn.ReLU(inplace)
self.inception_4e_double_3x3_1 = nn.Conv2d(
192, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
self.inception_4e_double_3x3_1_bn = nn.BatchNorm2d(256, affine=True)
self.inception_4e_relu_double_3x3_1 = nn.ReLU(inplace)
self.inception_4e_double_3x3_2 = nn.Conv2d(
256, 256, kernel_size=(3, 3), stride=(2, 2), padding=(1, 1))
self.inception_4e_double_3x3_2_bn = nn.BatchNorm2d(256, affine=True)
self.inception_4e_relu_double_3x3_2 = nn.ReLU(inplace)
self.inception_4e_pool = nn.MaxPool2d((3, 3),
stride=(2, 2),
dilation=(1, 1),
ceil_mode=True)
self.inception_5a_1x1 = nn.Conv2d(
1056, 352, kernel_size=(1, 1), stride=(1, 1))
self.inception_5a_1x1_bn = nn.BatchNorm2d(352, affine=True)
self.inception_5a_relu_1x1 = nn.ReLU(inplace)
self.inception_5a_3x3_reduce = nn.Conv2d(
1056, 192, kernel_size=(1, 1), stride=(1, 1))
self.inception_5a_3x3_reduce_bn = nn.BatchNorm2d(192, affine=True)
self.inception_5a_relu_3x3_reduce = nn.ReLU(inplace)
self.inception_5a_3x3 = nn.Conv2d(
192, 320, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
self.inception_5a_3x3_bn = nn.BatchNorm2d(320, affine=True)
self.inception_5a_relu_3x3 = nn.ReLU(inplace)
self.inception_5a_double_3x3_reduce = nn.Conv2d(
1056, 160, kernel_size=(1, 1), stride=(1, 1))
self.inception_5a_double_3x3_reduce_bn = nn.BatchNorm2d(
160, affine=True)
self.inception_5a_relu_double_3x3_reduce = nn.ReLU(inplace)
self.inception_5a_double_3x3_1 = nn.Conv2d(
160, 224, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
self.inception_5a_double_3x3_1_bn = nn.BatchNorm2d(224, affine=True)
self.inception_5a_relu_double_3x3_1 = nn.ReLU(inplace)
self.inception_5a_double_3x3_2 = nn.Conv2d(
224, 224, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
self.inception_5a_double_3x3_2_bn = nn.BatchNorm2d(224, affine=True)
self.inception_5a_relu_double_3x3_2 = nn.ReLU(inplace)
self.inception_5a_pool = nn.AvgPool2d(
3, stride=1, padding=1, ceil_mode=True, count_include_pad=True)
self.inception_5a_pool_proj = nn.Conv2d(
1056, 128, kernel_size=(1, 1), stride=(1, 1))
self.inception_5a_pool_proj_bn = nn.BatchNorm2d(128, affine=True)
self.inception_5a_relu_pool_proj = nn.ReLU(inplace)
self.inception_5b_1x1 = nn.Conv2d(
1024, 352, kernel_size=(1, 1), stride=(1, 1))
self.inception_5b_1x1_bn = nn.BatchNorm2d(352, affine=True)
self.inception_5b_relu_1x1 = nn.ReLU(inplace)
self.inception_5b_3x3_reduce = nn.Conv2d(
1024, 192, kernel_size=(1, 1), stride=(1, 1))
self.inception_5b_3x3_reduce_bn = nn.BatchNorm2d(192, affine=True)
self.inception_5b_relu_3x3_reduce = nn.ReLU(inplace)
self.inception_5b_3x3 = nn.Conv2d(
192, 320, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
self.inception_5b_3x3_bn = nn.BatchNorm2d(320, affine=True)
self.inception_5b_relu_3x3 = nn.ReLU(inplace)
self.inception_5b_double_3x3_reduce = nn.Conv2d(
1024, 192, kernel_size=(1, 1), stride=(1, 1))
self.inception_5b_double_3x3_reduce_bn = nn.BatchNorm2d(
192, affine=True)
self.inception_5b_relu_double_3x3_reduce = nn.ReLU(inplace)
self.inception_5b_double_3x3_1 = nn.Conv2d(
192, 224, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
self.inception_5b_double_3x3_1_bn = nn.BatchNorm2d(224, affine=True)
self.inception_5b_relu_double_3x3_1 = nn.ReLU(inplace)
self.inception_5b_double_3x3_2 = nn.Conv2d(
224, 224, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
self.inception_5b_double_3x3_2_bn = nn.BatchNorm2d(224, affine=True)
self.inception_5b_relu_double_3x3_2 = nn.ReLU(inplace)
self.inception_5b_pool = nn.MaxPool2d((3, 3),
stride=(1, 1),
padding=(1, 1),
dilation=(1, 1),
ceil_mode=True)
self.inception_5b_pool_proj = nn.Conv2d(
1024, 128, kernel_size=(1, 1), stride=(1, 1))
self.inception_5b_pool_proj_bn = nn.BatchNorm2d(128, affine=True)
self.inception_5b_relu_pool_proj = nn.ReLU(inplace)
self.num_classes = num_classes
if num_classes > 0:
self.last_linear = nn.Linear(1024, num_classes)
self.default_pretrained_model_path = model_urls[
self.__class__.__name__]