super-gradients/super_gradients/training/models/classification_models/mobilenetv2.py

"""
This is a PyTorch implementation of MobileNetV2 architecture as described in the paper:
Inverted Residuals and Linear Bottlenecks: Mobile Networks for Classification, Detection and Segmentation.
https://arxiv.org/pdf/1801.04381

Code taken from https://github.com/tonylins/pytorch-mobilenet-v2
License: Apache Version 2.0, January 2004 http://www.apache.org/licenses/

Pre-trained ImageNet model: 'deci-model-repository/mobilenet_v2/ckpt_best.pth'
"""
import numpy as np
import torch
import torch.nn as nn
import math
from super_gradients.training.models.sg_module import SgModule


def conv_bn(inp, oup, stride):
    return nn.Sequential(
        nn.Conv2d(inp, oup, 3, stride, 1, bias=False),
        nn.BatchNorm2d(oup),
        nn.ReLU6(inplace=True)
    )


def conv_1x1_bn(inp, oup):
    return nn.Sequential(
        nn.Conv2d(inp, oup, 1, 1, 0, bias=False),
        nn.BatchNorm2d(oup),
        nn.ReLU6(inplace=True)
    )


def make_divisible(x, divisible_by=8):
    import numpy as np
    return int(np.ceil(x * 1. / divisible_by) * divisible_by)


class InvertedResidual(nn.Module):
    def __init__(self, inp, oup, stride, expand_ratio, grouped_conv_size=1):
        """
        :param inp: number of input channels
        :param oup: number of output channels
        :param stride: conv stride
        :param expand_ratio: expansion ratio of the hidden layer after pointwise conv
        :grouped_conv_size: number of channels per grouped convolution, for depth-wise-separable convolution, use grouped_conv_size=1
        """
        super(InvertedResidual, self).__init__()
        self.stride = stride
        assert stride in [1, 2]

        hidden_dim = int(inp * expand_ratio)
        groups = int(hidden_dim / grouped_conv_size)
        self.use_res_connect = self.stride == 1 and inp == oup

        if expand_ratio == 1:
            self.conv = nn.Sequential(
                # dw
                nn.Conv2d(hidden_dim, hidden_dim, 3, stride, 1, groups=groups, bias=False),
                nn.BatchNorm2d(hidden_dim),
                nn.ReLU6(inplace=True),
                # pw-linear
                nn.Conv2d(hidden_dim, oup, 1, 1, 0, bias=False),
                nn.BatchNorm2d(oup),
            )
        else:
            self.conv = nn.Sequential(
                # pw
                nn.Conv2d(inp, hidden_dim, 1, 1, 0, bias=False),
                nn.BatchNorm2d(hidden_dim),
                nn.ReLU6(inplace=True),
                # dw
                nn.Conv2d(hidden_dim, hidden_dim, 3, stride, 1, groups=groups, bias=False),
                nn.BatchNorm2d(hidden_dim),
                nn.ReLU6(inplace=True),
                # pw-linear
                nn.Conv2d(hidden_dim, oup, 1, 1, 0, bias=False),
                nn.BatchNorm2d(oup),
            )

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


class MobileNetV2(SgModule):
    def __init__(self, num_classes, width_mult=1., structure=None, backbone_mode: bool = False,
                 grouped_conv_size=1) -> object:
        super(MobileNetV2, self).__init__()
        block = InvertedResidual
        input_channel = 32
        last_channel = 1280
        # IF STRUCTURE IS NONE - USE THE DEFAULT STRUCTURE NOTED
        #                                                  t, c,  n, s    stage-0 is the first conv_bn layer
        self.interverted_residual_setting = structure or [[1, 16, 1, 1],  # stage-1
                                                          [6, 24, 2, 2],  # stage-2
                                                          [6, 32, 3, 2],  # stage-3
                                                          [6, 64, 4, 2],  # stage-4
                                                          [6, 96, 3, 1],  # stage-5
                                                          [6, 160, 3, 2],  # stage-6
                                                          [6, 320, 1, 1]]  # stage-7
        #                                                                   stage-8  is the last_layer

        self.last_channel = make_divisible(last_channel * width_mult) if width_mult > 1.0 else last_channel
        self.features = [conv_bn(3, input_channel, 2)]
        # building inverted residual blocks
        for t, c, n, s in self.interverted_residual_setting:
            output_channel = make_divisible(c * width_mult) if t > 1 else c
            for i in range(n):
                if i == 0:
                    self.features.append(
                        block(input_channel, output_channel, s, expand_ratio=t, grouped_conv_size=grouped_conv_size))
                else:
                    self.features.append(
                        block(input_channel, output_channel, 1, expand_ratio=t, grouped_conv_size=grouped_conv_size))
                input_channel = output_channel
        # building last several layers
        self.features.append(conv_1x1_bn(input_channel, self.last_channel))
        # make it nn.Sequential
        self.features = nn.Sequential(*self.features)

        if backbone_mode:
            self.classifier = nn.Identity()
            self.connection_layers_input_channel_size = self._extract_connection_layers_input_channel_size()
        else:
            # building classifier
            self.classifier = nn.Linear(self.last_channel, num_classes)
        self._initialize_weights()

    def forward(self, x):
        x = self.features(x)
        x = x.mean(3).mean(2)
        x = self.classifier(x)
        return x

    def _extract_connection_layers_input_channel_size(self):
        """
        Extracts the number of channels out when using mobilenetV2 as yolo backbone
        """
        curr_layer_input = torch.rand(1, 3, 320, 320)  # input dims are used to extract number of channels
        layers_num_to_extract = [np.array(self.interverted_residual_setting)[:stage, 2].sum() for stage in [3, 5]]
        connection_layers_input_channel_size = []
        for layer_idx, feature in enumerate(self.features):
            curr_layer_input = feature(curr_layer_input)
            if layer_idx in layers_num_to_extract:
                connection_layers_input_channel_size.append(curr_layer_input.shape[1])
        connection_layers_input_channel_size.append(self.last_channel)
        connection_layers_input_channel_size.reverse()
        return connection_layers_input_channel_size

    def _initialize_weights(self):
        for m in self.modules():
            if isinstance(m, nn.Conv2d):
                n = m.kernel_size[0] * m.kernel_size[1] * m.out_channels
                m.weight.data.normal_(0, math.sqrt(2. / n))
                if m.bias is not None:
                    m.bias.data.zero_()
            elif isinstance(m, nn.BatchNorm2d):
                m.weight.data.fill_(1)
                m.bias.data.zero_()
            elif isinstance(m, nn.Linear):
                n = m.weight.size(1)
                m.weight.data.normal_(0, 0.01)
                m.bias.data.zero_()


def mobile_net_v2(arch_params):
    """
    :param arch_params: HpmStruct
        must contain: 'num_classes': int
    :return: MobileNetV2: nn.Module
    """
    return MobileNetV2(num_classes=arch_params.num_classes, width_mult=1., structure=None)


def mobile_net_v2_135(arch_params):
    """
    This Model achieves 75.73% on Imagenet - similar to Resnet50
    :param arch_params: HpmStruct
        must contain: 'num_classes': int
    :return: MobileNetV2: nn.Module
    """

    return MobileNetV2(num_classes=arch_params.num_classes, width_mult=1.35, structure=None)


def custom_mobile_net_v2(arch_params):
    """
    :param arch_params: HpmStruct
        must contain:
            'num_classes': int
            'width_mult': float
            'structure' : list. specify the mobilenetv2 architecture
    :return: MobileNetV2: nn.Module
    """

    return MobileNetV2(num_classes=arch_params.num_classes, width_mult=arch_params.width_mult,
                       structure=arch_params.structure)