Skip to content
Snippets Groups Projects
Commit e08a1fb7 authored by André Anjos's avatar André Anjos :speech_balloon:
Browse files

[modelling,utils] Avoid method names with underscore to avoid sphinx errors

parent ba9be5b4
No related branches found
No related tags found
1 merge request!12Streamlining
Pipeline #38679 passed
Stage: build
Hide whitespace changes
Inline Side-by-side
bob/ip/binseg/modeling/backbones/resnet.py
+ 20
− 20
View file @ e08a1fb7
...@@ -15,27 +15,27 @@ model_urls = { ...@@ -15,27 +15,27 @@ model_urls = {
} }
def _conv3x3(in_planes, out_planes, stride=1): def conv3x3(in_planes, out_planes, stride=1):
"""3x3 convolution with padding""" """3x3 convolution with padding"""
return nn.Conv2d( return nn.Conv2d(
in_planes, out_planes, kernel_size=3, stride=stride, padding=1, bias=False, in_planes, out_planes, kernel_size=3, stride=stride, padding=1, bias=False,
) )
def _conv1x1(in_planes, out_planes, stride=1): def conv1x1(in_planes, out_planes, stride=1):
"""1x1 convolution""" """1x1 convolution"""
return nn.Conv2d(in_planes, out_planes, kernel_size=1, stride=stride, bias=False) return nn.Conv2d(in_planes, out_planes, kernel_size=1, stride=stride, bias=False)
class _BasicBlock(nn.Module): class BasicBlock(nn.Module):
expansion = 1 expansion = 1
def __init__(self, inplanes, planes, stride=1, downsample=None): def __init__(self, inplanes, planes, stride=1, downsample=None):
super(_BasicBlock, self).__init__() super(BasicBlock, self).__init__()
self.conv1 = _conv3x3(inplanes, planes, stride) self.conv1 = conv3x3(inplanes, planes, stride)
self.bn1 = nn.BatchNorm2d(planes) self.bn1 = nn.BatchNorm2d(planes)
self.relu = nn.ReLU(inplace=True) self.relu = nn.ReLU(inplace=True)
self.conv2 = _conv3x3(planes, planes) self.conv2 = conv3x3(planes, planes)
self.bn2 = nn.BatchNorm2d(planes) self.bn2 = nn.BatchNorm2d(planes)
self.downsample = downsample self.downsample = downsample
self.stride = stride self.stride = stride
...@@ -59,16 +59,16 @@ class _BasicBlock(nn.Module): ...@@ -59,16 +59,16 @@ class _BasicBlock(nn.Module):
return out return out
class _Bottleneck(nn.Module): class Bottleneck(nn.Module):
expansion = 4 expansion = 4
def __init__(self, inplanes, planes, stride=1, downsample=None): def __init__(self, inplanes, planes, stride=1, downsample=None):
super(_Bottleneck, self).__init__() super(Bottleneck, self).__init__()
self.conv1 = _conv1x1(inplanes, planes) self.conv1 = conv1x1(inplanes, planes)
self.bn1 = nn.BatchNorm2d(planes) self.bn1 = nn.BatchNorm2d(planes)
self.conv2 = _conv3x3(planes, planes, stride) self.conv2 = conv3x3(planes, planes, stride)
self.bn2 = nn.BatchNorm2d(planes) self.bn2 = nn.BatchNorm2d(planes)
self.conv3 = _conv1x1(planes, planes * self.expansion) self.conv3 = conv1x1(planes, planes * self.expansion)
self.bn3 = nn.BatchNorm2d(planes * self.expansion) self.bn3 = nn.BatchNorm2d(planes * self.expansion)
self.relu = nn.ReLU(inplace=True) self.relu = nn.ReLU(inplace=True)
self.downsample = downsample self.downsample = downsample
...@@ -141,16 +141,16 @@ class ResNet(nn.Module): ...@@ -141,16 +141,16 @@ class ResNet(nn.Module):
# This improves the model by 0.2~0.3% according to https://arxiv.org/abs/1706.02677 # This improves the model by 0.2~0.3% according to https://arxiv.org/abs/1706.02677
if zero_init_residual: if zero_init_residual:
for m in self.modules(): for m in self.modules():
if isinstance(m, _Bottleneck): if isinstance(m, Bottleneck):
nn.init.constant_(m.bn3.weight, 0) nn.init.constant_(m.bn3.weight, 0)
elif isinstance(m, _BasicBlock): elif isinstance(m, BasicBlock):
nn.init.constant_(m.bn2.weight, 0) nn.init.constant_(m.bn2.weight, 0)
def _make_layer(self, block, planes, blocks, stride=1): def _make_layer(self, block, planes, blocks, stride=1):
downsample = None downsample = None
if stride != 1 or self.inplanes != planes * block.expansion: if stride != 1 or self.inplanes != planes * block.expansion:
downsample = nn.Sequential( downsample = nn.Sequential(
_conv1x1(self.inplanes, planes * block.expansion, stride), conv1x1(self.inplanes, planes * block.expansion, stride),
nn.BatchNorm2d(planes * block.expansion), nn.BatchNorm2d(planes * block.expansion),
) )
...@@ -179,7 +179,7 @@ def resnet18(pretrained=False, **kwargs): ...@@ -179,7 +179,7 @@ def resnet18(pretrained=False, **kwargs):
Args: Args:
pretrained (bool): If True, returns a model pre-trained on ImageNet pretrained (bool): If True, returns a model pre-trained on ImageNet
""" """
model = ResNet(_BasicBlock, [2, 2, 2, 2], **kwargs) model = ResNet(BasicBlock, [2, 2, 2, 2], **kwargs)
if pretrained: if pretrained:
model.load_state_dict(model_zoo.load_url(model_urls["resnet18"])) model.load_state_dict(model_zoo.load_url(model_urls["resnet18"]))
return model return model
...@@ -190,7 +190,7 @@ def resnet34(pretrained=False, **kwargs): ...@@ -190,7 +190,7 @@ def resnet34(pretrained=False, **kwargs):
Args: Args:
pretrained (bool): If True, returns a model pre-trained on ImageNet pretrained (bool): If True, returns a model pre-trained on ImageNet
""" """
model = ResNet(_BasicBlock, [3, 4, 6, 3], **kwargs) model = ResNet(BasicBlock, [3, 4, 6, 3], **kwargs)
if pretrained: if pretrained:
model.load_state_dict(model_zoo.load_url(model_urls["resnet34"])) model.load_state_dict(model_zoo.load_url(model_urls["resnet34"]))
return model return model
...@@ -201,7 +201,7 @@ def resnet50(pretrained=False, **kwargs): ...@@ -201,7 +201,7 @@ def resnet50(pretrained=False, **kwargs):
Args: Args:
pretrained (bool): If True, returns a model pre-trained on ImageNet pretrained (bool): If True, returns a model pre-trained on ImageNet
""" """
model = ResNet(_Bottleneck, [3, 4, 6, 3], **kwargs) model = ResNet(Bottleneck, [3, 4, 6, 3], **kwargs)
if pretrained: if pretrained:
model.load_state_dict(model_zoo.load_url(model_urls["resnet50"])) model.load_state_dict(model_zoo.load_url(model_urls["resnet50"]))
return model return model
...@@ -212,7 +212,7 @@ def shaperesnet50(pretrained=False, **kwargs): ...@@ -212,7 +212,7 @@ def shaperesnet50(pretrained=False, **kwargs):
Args: Args:
pretrained (bool): If True, returns a model pre-trained on ImageNet pretrained (bool): If True, returns a model pre-trained on ImageNet
""" """
model = ResNet(_Bottleneck, [3, 4, 6, 3], **kwargs) model = ResNet(Bottleneck, [3, 4, 6, 3], **kwargs)
if pretrained: if pretrained:
model.load_state_dict( model.load_state_dict(
model_zoo.load_url( model_zoo.load_url(
...@@ -227,7 +227,7 @@ def resnet101(pretrained=False, **kwargs): ...@@ -227,7 +227,7 @@ def resnet101(pretrained=False, **kwargs):
Args: Args:
pretrained (bool): If True, returns a model pre-trained on ImageNet pretrained (bool): If True, returns a model pre-trained on ImageNet
""" """
model = ResNet(_Bottleneck, [3, 4, 23, 3], **kwargs) model = ResNet(Bottleneck, [3, 4, 23, 3], **kwargs)
if pretrained: if pretrained:
model.load_state_dict(model_zoo.load_url(model_urls["resnet101"])) model.load_state_dict(model_zoo.load_url(model_urls["resnet101"]))
return model return model
...@@ -238,7 +238,7 @@ def resnet152(pretrained=False, **kwargs): ...@@ -238,7 +238,7 @@ def resnet152(pretrained=False, **kwargs):
Args: Args:
pretrained (bool): If True, returns a model pre-trained on ImageNet pretrained (bool): If True, returns a model pre-trained on ImageNet
""" """
model = ResNet(_Bottleneck, [3, 8, 36, 3], **kwargs) model = ResNet(Bottleneck, [3, 8, 36, 3], **kwargs)
if pretrained: if pretrained:
model.load_state_dict(model_zoo.load_url(model_urls["resnet152"])) model.load_state_dict(model_zoo.load_url(model_urls["resnet152"]))
return model return model
bob/ip/binseg/modeling/backbones/vgg.py
+ 9
− 9
View file @ e08a1fb7
...@@ -52,7 +52,7 @@ class VGG(nn.Module): ...@@ -52,7 +52,7 @@ class VGG(nn.Module):
nn.init.constant_(m.bias, 0) nn.init.constant_(m.bias, 0)
def _make_layers(cfg, batch_norm=False): def make_layers(cfg, batch_norm=False):
layers = [] layers = []
in_channels = 3 in_channels = 3
for v in cfg: for v in cfg:
...@@ -124,7 +124,7 @@ def vgg11(pretrained=False, **kwargs): ...@@ -124,7 +124,7 @@ def vgg11(pretrained=False, **kwargs):
""" """
if pretrained: if pretrained:
kwargs["init_weights"] = False kwargs["init_weights"] = False
model = VGG(_make_layers(_cfg["A"]), **kwargs) model = VGG(make_layers(_cfg["A"]), **kwargs)
if pretrained: if pretrained:
model.load_state_dict(model_zoo.load_url(model_urls["vgg11"])) model.load_state_dict(model_zoo.load_url(model_urls["vgg11"]))
return model return model
...@@ -137,7 +137,7 @@ def vgg11_bn(pretrained=False, **kwargs): ...@@ -137,7 +137,7 @@ def vgg11_bn(pretrained=False, **kwargs):
""" """
if pretrained: if pretrained:
kwargs["init_weights"] = False kwargs["init_weights"] = False
model = VGG(_make_layers(_cfg["A"], batch_norm=True), **kwargs) model = VGG(make_layers(_cfg["A"], batch_norm=True), **kwargs)
if pretrained: if pretrained:
model.load_state_dict(model_zoo.load_url(model_urls["vgg11_bn"])) model.load_state_dict(model_zoo.load_url(model_urls["vgg11_bn"]))
return model return model
...@@ -150,7 +150,7 @@ def vgg13(pretrained=False, **kwargs): ...@@ -150,7 +150,7 @@ def vgg13(pretrained=False, **kwargs):
""" """
if pretrained: if pretrained:
kwargs["init_weights"] = False kwargs["init_weights"] = False
model = VGG(_make_layers(_cfg["B"]), **kwargs) model = VGG(make_layers(_cfg["B"]), **kwargs)
if pretrained: if pretrained:
model.load_state_dict(model_zoo.load_url(model_urls["vgg13"])) model.load_state_dict(model_zoo.load_url(model_urls["vgg13"]))
return model return model
...@@ -163,7 +163,7 @@ def vgg13_bn(pretrained=False, **kwargs): ...@@ -163,7 +163,7 @@ def vgg13_bn(pretrained=False, **kwargs):
""" """
if pretrained: if pretrained:
kwargs["init_weights"] = False kwargs["init_weights"] = False
model = VGG(_make_layers(_cfg["B"], batch_norm=True), **kwargs) model = VGG(make_layers(_cfg["B"], batch_norm=True), **kwargs)
if pretrained: if pretrained:
model.load_state_dict(model_zoo.load_url(model_urls["vgg13_bn"])) model.load_state_dict(model_zoo.load_url(model_urls["vgg13_bn"]))
return model return model
...@@ -176,7 +176,7 @@ def vgg16(pretrained=False, **kwargs): ...@@ -176,7 +176,7 @@ def vgg16(pretrained=False, **kwargs):
""" """
if pretrained: if pretrained:
kwargs["init_weights"] = False kwargs["init_weights"] = False
model = VGG(_make_layers(_cfg["D"]), **kwargs) model = VGG(make_layers(_cfg["D"]), **kwargs)
if pretrained: if pretrained:
model.load_state_dict(model_zoo.load_url(model_urls["vgg16"]), strict=False) model.load_state_dict(model_zoo.load_url(model_urls["vgg16"]), strict=False)
return model return model
...@@ -189,7 +189,7 @@ def vgg16_bn(pretrained=False, **kwargs): ...@@ -189,7 +189,7 @@ def vgg16_bn(pretrained=False, **kwargs):
""" """
if pretrained: if pretrained:
kwargs["init_weights"] = False kwargs["init_weights"] = False
model = VGG(_make_layers(_cfg["D"], batch_norm=True), **kwargs) model = VGG(make_layers(_cfg["D"], batch_norm=True), **kwargs)
if pretrained: if pretrained:
model.load_state_dict(model_zoo.load_url(model_urls["vgg16_bn"])) model.load_state_dict(model_zoo.load_url(model_urls["vgg16_bn"]))
return model return model
...@@ -202,7 +202,7 @@ def vgg19(pretrained=False, **kwargs): ...@@ -202,7 +202,7 @@ def vgg19(pretrained=False, **kwargs):
""" """
if pretrained: if pretrained:
kwargs["init_weights"] = False kwargs["init_weights"] = False
model = VGG(_make_layers(_cfg["E"]), **kwargs) model = VGG(make_layers(_cfg["E"]), **kwargs)
if pretrained: if pretrained:
model.load_state_dict(model_zoo.load_url(model_urls["vgg19"])) model.load_state_dict(model_zoo.load_url(model_urls["vgg19"]))
return model return model
...@@ -215,7 +215,7 @@ def vgg19_bn(pretrained=False, **kwargs): ...@@ -215,7 +215,7 @@ def vgg19_bn(pretrained=False, **kwargs):
""" """
if pretrained: if pretrained:
kwargs["init_weights"] = False kwargs["init_weights"] = False
model = VGG(_make_layers(_cfg["E"], batch_norm=True), **kwargs) model = VGG(make_layers(_cfg["E"], batch_norm=True), **kwargs)
if pretrained: if pretrained:
model.load_state_dict(model_zoo.load_url(model_urls["vgg19_bn"])) model.load_state_dict(model_zoo.load_url(model_urls["vgg19_bn"]))
return model return model
bob/ip/binseg/utils/model_zoo.py
+ 2
− 2
View file @ e08a1fb7
...@@ -35,7 +35,7 @@ modelurls = { ...@@ -35,7 +35,7 @@ modelurls = {
"""URLs of pre-trained models (backbones)""" """URLs of pre-trained models (backbones)"""
def _download_url_to_file(url, dst, hash_prefix, progress): def download_url_to_file(url, dst, hash_prefix, progress):
file_size = None file_size = None
u = urlopen(url) u = urlopen(url)
meta = u.info() meta = u.info()
...@@ -109,6 +109,6 @@ def cache_url(url, model_dir=None, progress=True): ...@@ -109,6 +109,6 @@ def cache_url(url, model_dir=None, progress=True):
hash_prefix = HASH_REGEX.search(filename) hash_prefix = HASH_REGEX.search(filename)
if hash_prefix is not None: if hash_prefix is not None:
hash_prefix = hash_prefix.group(1) hash_prefix = hash_prefix.group(1)
_download_url_to_file(url, cached_file, hash_prefix, progress=progress) download_url_to_file(url, cached_file, hash_prefix, progress=progress)
return cached_file return cached_file
0% or .
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment