Add ShuffleNet v2 (#849)

* Add ShuffleNet v2

Added 4 configurations: x0.5, x1, x1.5, x2
Add 2 pretrained models: x0.5, x1

* fix lint

* Change globalpool to torch.mean() call

docs/source/models.rst

@@ -11,6 +11,7 @@ architectures:
 -  `DenseNet`_
 -  `Inception`_ v3
 -  `GoogLeNet`_
+-  `ShuffleNet`_ v2
 
 You can construct a model with random weights by calling its constructor:
 
@@ -24,6 +25,7 @@ You can construct a model with random weights by calling its constructor:
     densenet = models.densenet161()
     inception = models.inception_v3()
     googlenet = models.googlenet()
+    shufflenet = models.shufflenetv2()
 
 We provide pre-trained models, using the PyTorch :mod:`torch.utils.model_zoo`.
 These can be constructed by passing ``pretrained=True``:
@@ -38,6 +40,7 @@ These can be constructed by passing ``pretrained=True``:
     densenet = models.densenet161(pretrained=True)
     inception = models.inception_v3(pretrained=True)
     googlenet = models.googlenet(pretrained=True)
+    shufflenet = models.shufflenetv2(pretrained=True)
 
 Instancing a pre-trained model will download its weights to a cache directory.
 This directory can be set using the `TORCH_MODEL_ZOO` environment variable. See
@@ -88,6 +91,7 @@ Densenet-201                      22.80           6.43
 Densenet-161                      22.35           6.20
 Inception v3                      22.55           6.44
 GoogleNet                         30.22           10.47
+ShuffleNet V2                     30.64           11.68
 ================================  =============   =============
 
 
@@ -98,6 +102,7 @@ GoogleNet                         30.22           10.47
 .. _DenseNet: https://arxiv.org/abs/1608.06993
 .. _Inception: https://arxiv.org/abs/1512.00567
 .. _GoogLeNet: https://arxiv.org/abs/1409.4842
+.. _ShuffleNet: https://arxiv.org/abs/1807.11164
 
 .. currentmodule:: torchvision.models
 
@@ -152,3 +157,8 @@ GoogLeNet
 
 .. autofunction:: googlenet
 
+ShuffleNet v2
+-------------
+
+.. autofunction:: shufflenet
+

torchvision/models/__init__.py

@@ -6,3 +6,4 @@ from .inception import *
 from .densenet import *
 from .googlenet import *
 from .mobilenet import *
+from .shufflenetv2 import *

torchvision/models/shufflenetv2.py

+import functools
+
+import torch
+import torch.nn as nn
+
+__all__ = ['ShuffleNetV2', 'shufflenetv2',
+           'shufflenetv2_x0_5', 'shufflenetv2_x1_0',
+           'shufflenetv2_x1_5', 'shufflenetv2_x2_0']
+
+model_urls = {
+    'shufflenetv2_x0.5':
+        'https://github.com/barrh/Shufflenet-v2-Pytorch/releases/download/v0.1.0/shufflenetv2_x0.5-f707e7126e.pt',
+    'shufflenetv2_x1.0':
+        'https://github.com/barrh/Shufflenet-v2-Pytorch/releases/download/v0.1.0/shufflenetv2_x1-5666bf0f80.pt',
+    'shufflenetv2_x1.5': None,
+    'shufflenetv2_x2.0': None,
+}
+
+
+def channel_shuffle(x, groups):
+    batchsize, num_channels, height, width = x.data.size()
+    channels_per_group = num_channels // groups
+
+    # reshape
+    x = x.view(batchsize, groups,
+               channels_per_group, height, width)
+
+    x = torch.transpose(x, 1, 2).contiguous()
+
+    # flatten
+    x = x.view(batchsize, -1, height, width)
+
+    return x
+
+
+class InvertedResidual(nn.Module):
+    def __init__(self, inp, oup, stride):
+        super(InvertedResidual, self).__init__()
+
+        if not (1 <= stride <= 3):
+            raise ValueError('illegal stride value')
+        self.stride = stride
+
+        branch_features = oup // 2
+        assert (self.stride != 1) or (inp == branch_features << 1)
+
+        pw_conv11 = functools.partial(nn.Conv2d, kernel_size=1, stride=1, padding=0, bias=False)
+        dw_conv33 = functools.partial(self.depthwise_conv,
+                                      kernel_size=3, stride=self.stride, padding=1)
+
+        if self.stride > 1:
+            self.branch1 = nn.Sequential(
+                dw_conv33(inp, inp),
+                nn.BatchNorm2d(inp),
+                pw_conv11(inp, branch_features),
+                nn.BatchNorm2d(branch_features),
+                nn.ReLU(inplace=True),
+            )
+
+        self.branch2 = nn.Sequential(
+            pw_conv11(inp if (self.stride > 1) else branch_features, branch_features),
+            nn.BatchNorm2d(branch_features),
+            nn.ReLU(inplace=True),
+            dw_conv33(branch_features, branch_features),
+            nn.BatchNorm2d(branch_features),
+            pw_conv11(branch_features, branch_features),
+            nn.BatchNorm2d(branch_features),
+            nn.ReLU(inplace=True),
+        )
+
+    @staticmethod
+    def depthwise_conv(i, o, kernel_size, stride=1, padding=0, bias=False):
+        return nn.Conv2d(i, o, kernel_size, stride, padding, bias=bias, groups=i)
+
+    def forward(self, x):
+        if self.stride == 1:
+            x1, x2 = x.chunk(2, dim=1)
+            out = torch.cat((x1, self.branch2(x2)), dim=1)
+        else:
+            out = torch.cat((self.branch1(x), self.branch2(x)), dim=1)
+
+        out = channel_shuffle(out, 2)
+
+        return out
+
+
+class ShuffleNetV2(nn.Module):
+    def __init__(self, num_classes=1000, input_size=224, width_mult=1):
+        super(ShuffleNetV2, self).__init__()
+
+        try:
+            self.stage_out_channels = self._getStages(float(width_mult))
+        except KeyError:
+            raise ValueError('width_mult {} is not supported'.format(width_mult))
+
+        input_channels = 3
+        output_channels = self.stage_out_channels[0]
+        self.conv1 = nn.Sequential(
+            nn.Conv2d(input_channels, output_channels, 3, 2, 1, bias=False),
+            nn.BatchNorm2d(output_channels),
+            nn.ReLU(inplace=True),
+        )
+        input_channels = output_channels
+
+        self.maxpool = nn.MaxPool2d(kernel_size=3, stride=2, padding=1)
+
+        stage_names = ['stage{}'.format(i) for i in [2, 3, 4]]
+        stage_repeats = [4, 8, 4]
+        for name, repeats, output_channels in zip(
+                stage_names, stage_repeats, self.stage_out_channels[1:]):
+            seq = [InvertedResidual(input_channels, output_channels, 2)]
+            for i in range(repeats - 1):
+                seq.append(InvertedResidual(output_channels, output_channels, 1))
+            setattr(self, name, nn.Sequential(*seq))
+            input_channels = output_channels
+
+        output_channels = self.stage_out_channels[-1]
+        self.conv5 = nn.Sequential(
+            nn.Conv2d(input_channels, output_channels, 1, 1, 0, bias=False),
+            nn.BatchNorm2d(output_channels),
+            nn.ReLU(inplace=True),
+        )
+
+        self.fc = nn.Linear(output_channels, num_classes)
+
+    def forward(self, x):
+        x = self.conv1(x)
+        x = self.maxpool(x)
+        x = self.stage2(x)
+        x = self.stage3(x)
+        x = self.stage4(x)
+        x = self.conv5(x)
+        x = x.mean([2, 3])  # globalpool
+        x = self.fc(x)
+        return x
+
+    @staticmethod
+    def _getStages(mult):
+        stages = {
+            '0.5': [24, 48, 96, 192, 1024],
+            '1.0': [24, 116, 232, 464, 1024],
+            '1.5': [24, 176, 352, 704, 1024],
+            '2.0': [24, 244, 488, 976, 2048],
+        }
+        return stages[str(mult)]
+
+
+def shufflenetv2(pretrained=False, num_classes=1000, input_size=224, width_mult=1, **kwargs):
+    model = ShuffleNetV2(num_classes=num_classes, input_size=input_size, width_mult=width_mult)
+
+    if pretrained:
+        # change width_mult to float
+        if isinstance(width_mult, int):
+            width_mult = float(width_mult)
+        model_type = ('_'.join([ShuffleNetV2.__name__, 'x' + str(width_mult)]))
+        try:
+            model_url = model_urls[model_type.lower()]
+        except KeyError:
+            raise ValueError('model {} is not support'.format(model_type))
+        if model_url is None:
+            raise NotImplementedError('pretrained {} is not supported'.format(model_type))
+        model.load_state_dict(torch.utils.model_zoo.load_url(model_url))
+
+    return model
+
+
+def shufflenetv2_x0_5(pretrained=False, num_classes=1000, input_size=224, **kwargs):
+    return shufflenetv2(pretrained, num_classes, input_size, 0.5)
+
+
+def shufflenetv2_x1_0(pretrained=False, num_classes=1000, input_size=224, **kwargs):
+    return shufflenetv2(pretrained, num_classes, input_size, 1)
+
+
+def shufflenetv2_x1_5(pretrained=False, num_classes=1000, input_size=224, **kwargs):
+    return shufflenetv2(pretrained, num_classes, input_size, 1.5)
+
+
+def shufflenetv2_x2_0(pretrained=False, num_classes=1000, input_size=224, **kwargs):
+    return shufflenetv2(pretrained, num_classes, input_size, 2)