import paddle.fluid.layers as F
import paddle.fluid.dygraph as nn
import paddle.fluid as fluid
import numpy as np
def conv3x3(in_planes, out_planes, stride=1):
return nn.Conv2D(in_planes, out_planes, filter_size=3, stride=stride, padding=1, bias_attr=False)
class ReLU(nn.Layer):
def __init__(self):
super(ReLU, self).__init__()
def forward(self, x):
return F.relu(x)
class Sigmoid(nn.Layer):
def __init__(self):
super(Sigmoid, self).__init__()
def forward(self, x):
return F.sigmoid(x)
class Bottleneck(nn.Layer):
expansion = 4
def __init__(self, inplanes, planes, stride=1, downsample=None):
super(Bottleneck, self).__init__()
self.conv1 = nn.Conv2D(inplanes, planes, filter_size=1, bias_attr=False)
self.bn1 = nn.BatchNorm(planes)
self.conv2 = nn.Conv2D(planes, planes, filter_size=3, stride=stride,
padding=1, bias_attr=False)
self.bn2 = nn.BatchNorm(planes)
self.conv3 = nn.Conv2D(planes, planes * 4, filter_size=1, bias_attr=False)
self.bn3 = nn.BatchNorm(planes * 4)
self.relu = ReLU()
self.downsample = downsample
self.stride = stride
def forward(self, x):
residual = x
out = self.conv1(x)
out = self.bn1(out)
out = self.relu(out)
out = self.conv2(out)
out = self.bn2(out)
out = self.relu(out)
out = self.conv3(out)
out = self.bn3(out)
if self.downsample is not None:
residual = self.downsample(x)
out += residual
out = self.relu(out)
return out
class ResNet(nn.Layer):
def __init__(self, block, layers, num_classes=1000):
self.inplanes = 128
super(ResNet, self).__init__()
self.conv1 = conv3x3(3, 64, stride=2)
self.bn1 = nn.BatchNorm(64)
self.relu1 = ReLU()
self.conv2 = conv3x3(64, 64)
self.bn2 = nn.BatchNorm(64)
self.relu2 = ReLU()
self.conv3 = conv3x3(64, 128)
self.bn3 = nn.BatchNorm(128)
self.relu3 = ReLU()
self.maxpool = nn.Pool2D(pool_padding=1,pool_size=3,pool_type="max",pool_stride=2)
self.layer1 = self._make_layer(block, 64, layers[0])
self.layer2 = self._make_layer(block, 128, layers[1], stride=2)
self.layer3 = self._make_layer(block, 256, layers[2], stride=2)
self.layer4 = self._make_layer(block, 512, layers[3], stride=2)
for m in self.sublayers():
if isinstance(m, nn.Conv2D):
n = m.weight.shape[0]*m.weight.shape[1]*m.weight.shape[2]
v = np.random.normal(loc=0.,scale=np.sqrt(2./n),size=m.weight.shape).astype('float32')
m.weight.set_value(v)
elif isinstance(m, nn.BatchNorm):
m.weight.set_value(np.ones(m.weight.shape).astype('float32'))
m.bias.set_value(np.zeros(m.bias.shape).astype('float32'))
def _make_layer(self, block, planes, blocks, stride=1):
downsample = None
if stride != 1 or self.inplanes != planes * block.expansion:
downsample = nn.Sequential(
nn.Conv2D(self.inplanes, planes * block.expansion,
filter_size=1, stride=stride, bias_attr=False),
nn.BatchNorm(planes * block.expansion),
)
layers = []
layers.append(block(self.inplanes, planes, stride, downsample))
self.inplanes = planes * block.expansion
for i in range(1, blocks):
layers.append(block(self.inplanes, planes))
return nn.Sequential(*layers)
def forward(self, x):
x = self.relu1(self.bn1(self.conv1(x)))
x = self.relu2(self.bn2(self.conv2(x)))
x = self.relu3(self.bn3(self.conv3(x)))
x = self.maxpool(x)
x = self.layer1(x)
x = self.layer2(x)
x = self.layer3(x)
x = self.layer4(x)
return x
def resnet50():
return ResNet(Bottleneck, [3,4,6,3])
