Source code for norse.torch.models.mobilenet

# adapted from https://github.com/pytorch/vision/blob/master/torchvision/models/mobilenet.py
# licensed under BSD 3-Clause License, see LICENSE.torchvision for license details
from torch import nn
from torch.hub import load_state_dict_from_url
from norse.torch.module.lif import LIF
from norse.torch.module.lift import Lift

__all__ = ["MobileNetV2", "mobilenet_v2"]


model_urls = {
    "mobilenet_v2": "https://download.pytorch.org/models/mobilenet_v2-b0353104.pth",
}


def _make_divisible(v, divisor):
    """
    This function is taken from the original tf repo.
    It ensures that all layers have a channel number that is divisible by 8
    It can be seen here:
    https://github.com/tensorflow/models/blob/master/research/slim/nets/mobilenet/mobilenet.py
    :param v:
    :param divisor:

    :return:
    """
    min_value = divisor

    new_v = max(min_value, int(v + divisor / 2) // divisor * divisor)
    # Make sure that round down does not go down by more than 10%.
    if new_v < 0.9 * v:
        new_v += divisor
    return new_v


class ConvBNReLU(nn.Sequential):
    def __init__(
        self, in_planes, out_planes, norm_layer, kernel_size=3, stride=1, groups=1
    ):
        padding = (kernel_size - 1) // 2

        super(ConvBNReLU, self).__init__(
            Lift(
                nn.Conv2d(
                    in_planes,
                    out_planes,
                    kernel_size,
                    stride,
                    padding,
                    groups=groups,
                    bias=False,
                )
            ),
            Lift(norm_layer(out_planes)),
            LIF(),
        )


class InvertedResidual(nn.Module):
    def __init__(self, inp, oup, stride, expand_ratio, norm_layer):
        super(InvertedResidual, self).__init__()
        self.stride = stride
        assert stride in [1, 2]

        hidden_dim = int(round(inp * expand_ratio))
        self.use_res_connect = self.stride == 1 and inp == oup

        layers = []
        if expand_ratio != 1:
            # pw
            layers.append(
                ConvBNReLU(inp, hidden_dim, kernel_size=1, norm_layer=norm_layer)
            )
        layers.extend(
            [
                # dw
                ConvBNReLU(
                    hidden_dim,
                    hidden_dim,
                    stride=stride,
                    groups=hidden_dim,
                    norm_layer=norm_layer,
                ),
                # pw-linear
                Lift(nn.Conv2d(hidden_dim, oup, 1, 1, 0, bias=False)),
                Lift(norm_layer(oup)),
            ]
        )
        self.conv = nn.Sequential(*layers)

    def forward(self, x):
        if self.use_res_connect:
            return x + self.conv(x)
        else:
            return self.conv(x)


[docs]class MobileNetV2(nn.Module): def __init__( self, num_classes=1000, width_mult=1.0, inverted_residual_setting=None, round_nearest=8, block=None, norm_layer=None, ): """ MobileNet V2 main class Args: num_classes (int): Number of classes width_mult (float): Width multiplier - adjusts number of channels in each layer by this amount inverted_residual_setting: Network structure round_nearest (int): Round the number of channels in each layer to be a multiple of this number Set to 1 to turn off rounding block: Module specifying inverted residual building block for mobilenet norm_layer: Module specifying the normalization layer to use """ super(MobileNetV2, self).__init__() if block is None: block = InvertedResidual if norm_layer is None: norm_layer = nn.BatchNorm2d input_channel = 32 last_channel = 1280 if inverted_residual_setting is None: inverted_residual_setting = [ # t, c, n, s [1, 16, 1, 1], [6, 24, 2, 2], [6, 32, 3, 2], [6, 64, 4, 2], [6, 96, 3, 1], [6, 160, 3, 2], [6, 320, 1, 1], ] # only check the first element, assuming user knows t,c,n,s are required if ( len(inverted_residual_setting) == 0 or len(inverted_residual_setting[0]) != 4 ): raise ValueError( "inverted_residual_setting should be non-empty " "or a 4-element list, got {}".format(inverted_residual_setting) ) # building first layer input_channel = _make_divisible(input_channel * width_mult, round_nearest) self.last_channel = _make_divisible( last_channel * max(1.0, width_mult), round_nearest ) features = [ConvBNReLU(3, input_channel, stride=2, norm_layer=norm_layer)] # building inverted residual blocks for t, c, n, s in inverted_residual_setting: output_channel = _make_divisible(c * width_mult, round_nearest) for i in range(n): stride = s if i == 0 else 1 features.append( block( input_channel, output_channel, stride, expand_ratio=t, norm_layer=norm_layer, ) ) input_channel = output_channel # building last several layers features.append( ConvBNReLU( input_channel, self.last_channel, kernel_size=1, norm_layer=norm_layer ) ) # make it nn.Sequential self.features = nn.Sequential(*features) self.pool = Lift(nn.AdaptiveAvgPool2d(1)) # building classifier self.classifier = nn.Sequential( Lift(nn.Dropout(0.2)), Lift(nn.Linear(self.last_channel, num_classes)), ) # weight initialization for m in self.modules(): if isinstance(m, nn.Conv2d): nn.init.kaiming_normal_(m.weight, mode="fan_out") assert m.bias is None elif isinstance(m, (nn.BatchNorm2d, nn.GroupNorm)): nn.init.ones_(m.weight) nn.init.zeros_(m.bias) elif isinstance(m, nn.Linear): nn.init.normal_(m.weight, 0, 0.01) nn.init.zeros_(m.bias) def _forward_impl(self, x): # This exists since TorchScript doesn't support inheritance, so the superclass method # (this one) needs to have a name other than `forward` that can be accessed in a subclass x, _ = self.features(x) # Cannot use "squeeze" as batch-size can be 1 => must use reshape with x.shape[0] # x = nn.functional.adaptive_avg_pool2d(x, 1).reshape(x.shape[1], -1) x = self.pool(x).reshape(x.shape[0], x.shape[1], -1) x = self.classifier(x) return x
[docs] def forward(self, x): return self._forward_impl(x)
[docs]def mobilenet_v2(pretrained=False, progress=True, **kwargs): """ Constructs a MobileNetV2 architecture from `"MobileNetV2: Inverted Residuals and Linear Bottlenecks" <https://arxiv.org/abs/1801.04381>`_. Args: pretrained (bool): If True, returns a model pre-trained on ImageNet progress (bool): If True, displays a progress bar of the download to stderr """ model = MobileNetV2(**kwargs) if pretrained: state_dict = load_state_dict_from_url( model_urls["mobilenet_v2"], progress=progress ) snn_state_dict = {} # compared to the ANN parameters, the modules are lifted # we modify the state dict accordingly here for key in state_dict: l = key.split(".") l.insert(-1, "lifted_module") new_key = ".".join(l) snn_state_dict[new_key] = state_dict[key] model.load_state_dict(snn_state_dict) return model