net.py

# -*- coding: utf-8 -*-

import torch
import torch.nn as nn
import torch.utils.checkpoint as checkpoint
from timm.models.layers import DropPath, to_2tuple, trunc_normal_
import numpy as np
import torch.nn.functional as F


class Decoder(nn.Module):
    def __init__(self):
        super(Decoder, self).__init__()
        self.conv0 = nn.Conv2d(64, 64, kernel_size=3, stride=1, padding=1)
        self.bn0   = nn.BatchNorm2d(64)
        self.conv1 = nn.Conv2d(64, 64, kernel_size=3, stride=1, padding=1)
        self.bn1   = nn.BatchNorm2d(64)
        self.conv2 = nn.Conv2d(64, 64, kernel_size=3, stride=1, padding=1)
        self.bn2   = nn.BatchNorm2d(64)
        self.conv3 = nn.Conv2d(64, 64, kernel_size=3, stride=1, padding=1)
        self.bn3   = nn.BatchNorm2d(64)

    def forward(self, input1, input2=[0,0,0,0]):
        out0 = F.relu(self.bn0(self.conv0(input1[0]+input2[0])), inplace=True)
        out0 = F.interpolate(out0, size=input1[1].size()[2:], mode='bilinear')
        out1 = F.relu(self.bn1(self.conv1(input1[1]+input2[1]+out0)), inplace=True)
        out1 = F.interpolate(out1, size=input1[2].size()[2:], mode='bilinear')
        out2 = F.relu(self.bn2(self.conv2(input1[2]+input2[2]+out1)), inplace=True)
        out2 = F.interpolate(out2, size=input1[3].size()[2:], mode='bilinear')
        out3 = F.relu(self.bn3(self.conv3(input1[3]+input2[3]+out2)), inplace=True)
        
        return out3



class ChannelAttention(nn.Module):
    def __init__(self, in_planes=192, ratio=16):
        super(ChannelAttention, self).__init__()
        self.avg_pool = nn.AdaptiveAvgPool2d(1)
        self.max_pool = nn.AdaptiveMaxPool2d(1)

        self.fc1   = nn.Conv2d(in_planes, in_planes // ratio, 1, bias=False)
        self.relu1 = nn.ReLU()
        self.fc2   = nn.Conv2d(in_planes // ratio, in_planes, 1, bias=False)

        self.sigmoid = nn.Sigmoid()

    def forward(self, x):
        avg_out = self.fc2(self.relu1(self.fc1(self.avg_pool(x))))
        max_out = self.fc2(self.relu1(self.fc1(self.max_pool(x))))
        out = avg_out + max_out
        return self.sigmoid(out)


class SpatialAttention(nn.Module):
    def __init__(self, kernel_size=7):
        super(SpatialAttention, self).__init__()
        assert kernel_size in (3, 7), 'kernel size must be 3 or 7'
        padding = 3 if kernel_size == 7 else 1
        self.conv1 = nn.Conv2d(2, 1, kernel_size, padding=padding, bias=False)

        self.sigmoid = nn.Sigmoid()

    def forward(self, x):
        avg_out    = torch.mean(x, dim=1, keepdim=True)
        max_out, _ = torch.max(x, dim=1, keepdim=True)
        x = torch.cat([avg_out, max_out], dim=1)
        x = self.conv1(x)

        return self.sigmoid(x)



class Attention(nn.Module):
    def __init__(self):
        super(Attention , self).__init__()
        
        self.att_c = ChannelAttention()
        self.att_s = SpatialAttention()

        self.fc_i1 = nn.Sequential(nn.Conv2d(64, 64, kernel_size=3, padding=1), nn.BatchNorm2d(64), nn.ReLU(inplace=True))
        self.fc_t1 = nn.Sequential(nn.Conv2d(64, 64, kernel_size=3, padding=1), nn.BatchNorm2d(64), nn.ReLU(inplace=True))

        self.fc_i2  = nn.Sequential(nn.Conv2d(192, 64, kernel_size=3, padding=1), nn.BatchNorm2d(64), nn.ReLU(inplace=True))
        self.fc_t2  = nn.Sequential(nn.Conv2d(192, 64, kernel_size=3, padding=1), nn.BatchNorm2d(64), nn.ReLU(inplace=True))
 

    def forward(self, i, t):

        sa_i = i.mul(self.att_s(i))
        sa_t = t.mul(self.att_s(t))

        i1 = self.fc_i1(i)
        t1 = self.fc_t1(t)
        
        mix2 = torch.cat([torch.cat([i1, torch.mul(i1, t1)], dim=1), t1], dim =1)
        ca = mix2.mul(self.att_c(mix2))        
        
        atti = self.fc_i2(ca) + sa_t
        attt = self.fc_t2(ca) + sa_i
        
        return atti, attt


class BasicConv(nn.Module):

    def __init__(self, in_planes, out_planes, kernel_size, stride=1, padding=0, dilation=1, groups=1, relu=True, bn=True, bias=False):
        super(BasicConv, self).__init__()
        self.out_channels = out_planes
        self.conv = nn.Conv2d(in_planes, out_planes, kernel_size=kernel_size, stride=stride, padding=padding, dilation=dilation, groups=groups, bias=bias)
        self.bn = nn.BatchNorm2d(out_planes,eps=1e-5, momentum=0.01, affine=True) if bn else None
        self.relu = nn.PReLU() if relu else None

    def forward(self, x):
        x = self.conv(x)
        if self.bn is not None:
            x = self.bn(x)
        if self.relu is not None:
            x = self.relu(x)
        return x


class LightRFB(nn.Module):
    def __init__(self, c_in=128, c_out=64):
        super(LightRFB, self).__init__()
        self.br2 = nn.Sequential(
            BasicConv(c_in, c_out, kernel_size=1,bias=False, bn=True, relu=True),

            BasicConv(c_out, c_out, kernel_size=3, dilation=1, padding=1, groups=c_out, bias=False,
                      relu=False),
        )
        self.br3 = nn.Sequential(
            BasicConv(c_out, c_out, kernel_size=3, dilation=1, padding=1, groups=c_out, bias=False,
                      bn=True,relu=False),
            BasicConv(c_out, c_out, kernel_size=1, dilation=1, bias=False,bn=True,relu=True),

            BasicConv(c_out, c_out, kernel_size=3, dilation=3, padding=3, groups=c_out, bias=False,
                      relu=False),
        )
        self.br4 = nn.Sequential(
            BasicConv(c_out, c_out, kernel_size=5, dilation=1, padding=2, groups=c_out, bias=False,
                      bn=True, relu=False),
            BasicConv(c_out, c_out, kernel_size=1, dilation=1, bias=False, bn=True, relu=True),

            BasicConv(c_out, c_out, kernel_size=3, dilation=5, padding=5, groups=c_out, bias=False,
                      relu=False),
        )
        self.br5 = nn.Sequential(
            BasicConv(c_out, c_out, kernel_size=7, dilation=1, padding=3, groups=c_out, bias=False,
                      bn=True, relu=False),
            BasicConv(c_out, c_out, kernel_size=1, dilation=1, bias=False, bn=True, relu=True),

            BasicConv(c_out, c_out, kernel_size=3, dilation=7, padding=7, groups=c_out, bias=False,
                      relu=False),
        )


        self.conv1b = nn.Conv2d(64, 64, kernel_size=3, stride=1, padding=1)
        self.bn1b   = nn.BatchNorm2d(64)
        self.conv2b = nn.Conv2d(64, 64, kernel_size=3, stride=1, padding=1)
        self.bn2b   = nn.BatchNorm2d(64)
        self.conv3b = nn.Conv2d(64, 64, kernel_size=3, stride=1, padding=1)
        self.bn3b   = nn.BatchNorm2d(64)
        self.conv4b = nn.Conv2d(64, 64, kernel_size=3, stride=1, padding=1)
        self.bn4b   = nn.BatchNorm2d(64)

        self.conv1d = nn.Conv2d(64, 64, kernel_size=3, stride=1, padding=1)
        self.bn1d   = nn.BatchNorm2d(64)
        self.conv2d = nn.Conv2d(64, 64, kernel_size=3, stride=1, padding=1)
        self.bn2d   = nn.BatchNorm2d(64)
        self.conv3d = nn.Conv2d(64, 64, kernel_size=3, stride=1, padding=1)
        self.bn3d   = nn.BatchNorm2d(64)
        self.conv4d = nn.Conv2d(64, 64, kernel_size=3, stride=1, padding=1)
        self.bn4d   = nn.BatchNorm2d(64)


    def forward(self, x):
        
        out2=self.br2(x)

        x3=self.br3(out2)
        out3 = F.max_pool2d(x3, kernel_size=2, stride=2)

        x4=self.br4(out3)
        out4 = F.max_pool2d(x4, kernel_size=2, stride=2)

        x5=self.br5(out4)
        out5 = F.max_pool2d(x5, kernel_size=2, stride=2)


        out1b = F.relu(self.bn1b(self.conv1b(out2)), inplace=True)
        out2b = F.relu(self.bn2b(self.conv2b(out3)), inplace=True)
        out3b = F.relu(self.bn3b(self.conv3b(out4)), inplace=True)
        out4b = F.relu(self.bn4b(self.conv4b(out5)), inplace=True)

        out1d = F.relu(self.bn1d(self.conv1d(out2)), inplace=True)
        out2d = F.relu(self.bn2d(self.conv2d(out3)), inplace=True)
        out3d = F.relu(self.bn3d(self.conv3d(out4)), inplace=True)
        out4d = F.relu(self.bn4d(self.conv4d(out5)), inplace=True)
        
        return (out4b, out3b, out2b, out1b), (out4d, out3d, out2d, out1d)



class Mlp(nn.Module):
    def __init__(self, in_features, hidden_features=None, out_features=None, act_layer=nn.GELU, drop=0.):
        super().__init__()
        out_features = out_features or in_features
        hidden_features = hidden_features or in_features
        self.fc1 = nn.Linear(in_features, hidden_features)
        self.act = act_layer()
        self.fc2 = nn.Linear(hidden_features, out_features)
        self.drop = nn.Dropout(drop)

    def forward(self, x):
        x = self.fc1(x)
        x = self.act(x)
        x = self.drop(x)
        x = self.fc2(x)
        x = self.drop(x)
        return x


def window_partition(x, window_size):
    """
    Args:
        x: (B, H, W, C)
        window_size (int): window size

    Returns:
        windows: (num_windows*B, window_size, window_size, C) 堆叠到一起形成一个长条
    """
    B, H, W, C = x.shape
    x = x.view(B, H // window_size, window_size, W // window_size, window_size, C)
    windows = x.permute(0, 1, 3, 2, 4, 5).contiguous().view(-1, window_size, window_size, C)
    return windows


def window_reverse(windows, window_size, H, W):
    """
    Args:
        windows: (num_windows*B, window_size, window_size, C)
        window_size (int): Window size
        H (int): Height of image
        W (int): Width of image

    Returns:
        x: (B, H, W, C)
    """
    B = int(windows.shape[0] / (H * W / window_size / window_size))
    x = windows.view(B, H // window_size, W // window_size, window_size, window_size, -1)
    x = x.permute(0, 1, 3, 2, 4, 5).contiguous().view(B, H, W, -1)
    return x


class WindowAttention(nn.Module):
    r""" Window based multi-head self attention (W-MSA) module with relative position bias.
    It supports both of shifted and non-shifted window.

    Args:
        dim (int): Number of input channels.
        window_size (tuple[int]): The height and width of the window.
        num_heads (int): Number of attention heads.
        qkv_bias (bool, optional):  If True, add a learnable bias to query, key, value. Default: True
        qk_scale (float | None, optional): Override default qk scale of head_dim ** -0.5 if set
        attn_drop (float, optional): Dropout ratio of attention weight. Default: 0.0
        proj_drop (float, optional): Dropout ratio of output. Default: 0.0
    """

    def __init__(self, dim, window_size, num_heads, qkv_bias=True, qk_scale=None, attn_drop=0., proj_drop=0.):

        super().__init__()
        self.dim = dim
        self.window_size = window_size  # Wh, Ww
        self.num_heads = num_heads
        head_dim = dim // num_heads # 每一个头的通道维数
        self.scale = qk_scale or head_dim ** -0.5

        # define a parameter table of relative position bias
        self.relative_position_bias_table = nn.Parameter(
            torch.zeros((2 * window_size[0] - 1) * (2 * window_size[1] - 1), num_heads))  # 2*Wh-1 * 2*Ww-1, nH

        # get pair-wise relative position index for each token inside the window
        coords_h = torch.arange(self.window_size[0])
        coords_w = torch.arange(self.window_size[1])

        coords = torch.stack(torch.meshgrid([coords_h, coords_w]))  # 2, Wh, Ww
        coords_flatten = torch.flatten(coords, 1)  # 2, Wh*Ww
        relative_coords = coords_flatten[:, :, None] - coords_flatten[:, None, :]  # 2, Wh*Ww, Wh*Ww
        relative_coords = relative_coords.permute(1, 2, 0).contiguous()  # Wh*Ww, Wh*Ww, 2
        relative_coords[:, :, 0] += self.window_size[0] - 1  # shift to start from 0
        relative_coords[:, :, 1] += self.window_size[1] - 1
        relative_coords[:, :, 0] *= 2 * self.window_size[1] - 1
        relative_position_index = relative_coords.sum(-1)  # Wh*Ww, Wh*Ww
        self.register_buffer("relative_position_index", relative_position_index)

        self.qkv = nn.Linear(dim, dim * 3, bias=qkv_bias)
        self.attn_drop = nn.Dropout(attn_drop)
        self.proj = nn.Linear(dim, dim)
        self.proj_drop = nn.Dropout(proj_drop)

        trunc_normal_(self.relative_position_bias_table, std=.02)
        self.softmax = nn.Softmax(dim=-1)

    def forward(self, x, mask=None):
        """
        Args:
            x: input features with shape of (num_windows*B, N, C)
            mask: (0/-inf) mask with shape of (num_windows, Wh*Ww, Wh*Ww) or None
        """
        B_, N, C = x.shape
        qkv = self.qkv(x).reshape(B_, N, 3, self.num_heads, C // self.num_heads).permute(2, 0, 3, 1, 4)
        q, k, v = qkv[0], qkv[1], qkv[2]  # make torchscript happy (cannot use tensor as tuple)

        q = q * self.scale
        attn = (q @ k.transpose(-2, -1))

        relative_position_bias = self.relative_position_bias_table[self.relative_position_index.view(-1)].view(
            self.window_size[0] * self.window_size[1], self.window_size[0] * self.window_size[1], -1)  # Wh*Ww,Wh*Ww,nH
        relative_position_bias = relative_position_bias.permute(2, 0, 1).contiguous()  # nH, Wh*Ww, Wh*Ww
        attn = attn + relative_position_bias.unsqueeze(0)

        if mask is not None:
            nW = mask.shape[0]
            attn = attn.view(B_ // nW, nW, self.num_heads, N, N) + mask.unsqueeze(1).unsqueeze(0)
            attn = attn.view(-1, self.num_heads, N, N)
            attn = self.softmax(attn)
        else:
            attn = self.softmax(attn)

        attn = self.attn_drop(attn)

        x = (attn @ v).transpose(1, 2).reshape(B_, N, C)
        x = self.proj(x)
        x = self.proj_drop(x)
        return x

    def extra_repr(self) -> str:
        return f'dim={self.dim}, window_size={self.window_size}, num_heads={self.num_heads}'

    def flops(self, N):
        # calculate flops for 1 window with token length of N
        flops = 0
        # qkv = self.qkv(x)
        flops += N * self.dim * 3 * self.dim
        # attn = (q @ k.transpose(-2, -1))
        flops += self.num_heads * N * (self.dim // self.num_heads) * N
        #  x = (attn @ v)
        flops += self.num_heads * N * N * (self.dim // self.num_heads)
        # x = self.proj(x)
        flops += N * self.dim * self.dim
        return flops


class SwinTransformerBlock(nn.Module):
    r""" Swin Transformer Block.

    Args:
        dim (int): Number of input channels.
        input_resolution (tuple[int]): Input resulotion.
        num_heads (int): Number of attention heads.
        window_size (int): Window size.
        shift_size (int): Shift size for SW-MSA.
        mlp_ratio (float): Ratio of mlp hidden dim to embedding dim.
        qkv_bias (bool, optional): If True, add a learnable bias to query, key, value. Default: True
        qk_scale (float | None, optional): Override default qk scale of head_dim ** -0.5 if set.
        drop (float, optional): Dropout rate. Default: 0.0
        attn_drop (float, optional): Attention dropout rate. Default: 0.0
        drop_path (float, optional): Stochastic depth rate. Default: 0.0
        act_layer (nn.Module, optional): Activation layer. Default: nn.GELU
        norm_layer (nn.Module, optional): Normalization layer.  Default: nn.LayerNorm
    """

    def __init__(self, dim, input_resolution, num_heads, window_size=7, shift_size=0,
                 mlp_ratio=4., qkv_bias=True, qk_scale=None, drop=0., attn_drop=0., drop_path=0.,
                 act_layer=nn.GELU, norm_layer=nn.LayerNorm):
        super().__init__()
        self.dim = dim
        self.input_resolution = input_resolution
        self.num_heads = num_heads
        self.window_size = window_size
        self.shift_size = shift_size
        self.mlp_ratio = mlp_ratio
        if min(self.input_resolution) <= self.window_size:
            # if window size is larger than input resolution, we don't partition windows
            self.shift_size = 0
            self.window_size = min(self.input_resolution)
        assert 0 <= self.shift_size < self.window_size, "shift_size must in 0-window_size"

        self.norm1 = norm_layer(dim)
        self.attn = WindowAttention(
            dim, window_size=to_2tuple(self.window_size), num_heads=num_heads,
            qkv_bias=qkv_bias, qk_scale=qk_scale, attn_drop=attn_drop, proj_drop=drop)

        self.drop_path = DropPath(drop_path) if drop_path > 0. else nn.Identity()
        self.norm2 = norm_layer(dim)
        mlp_hidden_dim = int(dim * mlp_ratio)
        self.mlp = Mlp(in_features=dim, hidden_features=mlp_hidden_dim, act_layer=act_layer, drop=drop)

        if self.shift_size > 0:
            # calculate attention mask for SW-MSA
            H, W = self.input_resolution
            img_mask = torch.zeros((1, H, W, 1))  # 1 H W 1---Important！！！
            h_slices = (slice(0, -self.window_size),
                        slice(-self.window_size, -self.shift_size),
                        slice(-self.shift_size, None))
            w_slices = (slice(0, -self.window_size),
                        slice(-self.window_size, -self.shift_size),
                        slice(-self.shift_size, None))
            cnt = 0
            for h in h_slices:
                for w in w_slices:
                    img_mask[:, h, w, :] = cnt
                    cnt += 1

            mask_windows = window_partition(img_mask, self.window_size)  # nW, window_size, window_size, 1
            mask_windows = mask_windows.view(-1, self.window_size * self.window_size)
            attn_mask = mask_windows.unsqueeze(1) - mask_windows.unsqueeze(2)
            attn_mask = attn_mask.masked_fill(attn_mask != 0, float(-100.0)).masked_fill(attn_mask == 0, float(0.0))
        else:
            attn_mask = None

        self.register_buffer("attn_mask", attn_mask)

    def forward(self, x):
        # 输入此的x是整图
        H, W = self.input_resolution
        B, L, C = x.shape
        assert L == H * W, "input feature has wrong size"

        shortcut = x
        x = self.norm1(x)
        x = x.view(B, H, W, C)

        # cyclic shift
        if self.shift_size > 0:

            shifted_x = torch.roll(x, shifts=(-self.shift_size, -self.shift_size), dims=(1, 2))
        else:
            shifted_x = x

        # partition windows
        x_windows = window_partition(shifted_x, self.window_size)  # nW*B, window_size, window_size, C
        x_windows = x_windows.view(-1, self.window_size * self.window_size, C)  # nW*B, window_size*window_size, C

        # W-MSA/SW-MSA
        attn_windows = self.attn(x_windows, mask=self.attn_mask)  # nW*B, window_size*window_size, C

        # merge windows
        attn_windows = attn_windows.view(-1, self.window_size, self.window_size, C)
        shifted_x = window_reverse(attn_windows, self.window_size, H, W)  # B H' W' C

        # reverse cyclic shift
        if self.shift_size > 0:
            x = torch.roll(shifted_x, shifts=(self.shift_size, self.shift_size), dims=(1, 2))
        else:
            x = shifted_x
        x = x.view(B, H * W, C)

        # FFN
        x = shortcut + self.drop_path(x)
        x = x + self.drop_path(self.mlp(self.norm2(x)))
        # print('FFN',x.shape)
        return x

    def extra_repr(self) -> str:
        return f"dim={self.dim}, input_resolution={self.input_resolution}, num_heads={self.num_heads}, " \
               f"window_size={self.window_size}, shift_size={self.shift_size}, mlp_ratio={self.mlp_ratio}"

    def flops(self):
        flops = 0
        H, W = self.input_resolution
        # norm1
        flops += self.dim * H * W
        # W-MSA/SW-MSA
        nW = H * W / self.window_size / self.window_size
        flops += nW * self.attn.flops(self.window_size * self.window_size)
        # mlp
        flops += 2 * H * W * self.dim * self.dim * self.mlp_ratio
        # norm2
        flops += self.dim * H * W
        return flops


class PatchMerging(nn.Module):
    r""" Patch Merging Layer.

    Args:
        input_resolution (tuple[int]): Resolution of input feature.
        dim (int): Number of input channels.
        norm_layer (nn.Module, optional): Normalization layer.  Default: nn.LayerNorm
    """

    def __init__(self, input_resolution, dim, norm_layer=nn.LayerNorm):
        super().__init__()
        self.input_resolution = input_resolution
        self.dim = dim
        self.reduction = nn.Linear(4 * dim, 2 * dim, bias=False)
        self.norm = norm_layer(4 * dim)

    def forward(self, x):
        """
        x: B, H*W, C
        """
        H, W = self.input_resolution
        B, L, C = x.shape
        assert L == H * W, "input feature has wrong size"
        assert H % 2 == 0 and W % 2 == 0, f"x size ({H}*{W}) are not even."

        x = x.view(B, H, W, C)

        x0 = x[:, 0::2, 0::2, :]  # B H/2 W/2 C
        x1 = x[:, 1::2, 0::2, :]  # B H/2 W/2 C
        x2 = x[:, 0::2, 1::2, :]  # B H/2 W/2 C
        x3 = x[:, 1::2, 1::2, :]  # B H/2 W/2 C
        x = torch.cat([x0, x1, x2, x3], -1)  # B H/2 W/2 4*C
        x = x.view(B, -1, 4 * C)  # B H/2*W/2 4*C

        x = self.norm(x)
        x = self.reduction(x)

        return x

    def extra_repr(self) -> str:
        return f"input_resolution={self.input_resolution}, dim={self.dim}"

    def flops(self):
        H, W = self.input_resolution
        flops = H * W * self.dim
        flops += (H // 2) * (W // 2) * 4 * self.dim * 2 * self.dim
        return flops


class BasicLayer(nn.Module):
    """ A basic Swin Transformer layer for one stage.

    Args:
        dim (int): Number of input channels.
        input_resolution (tuple[int]): Input resolution.
        depth (int): Number of blocks.
        num_heads (int): Number of attention heads.
        window_size (int): Local window size.
        mlp_ratio (float): Ratio of mlp hidden dim to embedding dim.
        qkv_bias (bool, optional): If True, add a learnable bias to query, key, value. Default: True
        qk_scale (float | None, optional): Override default qk scale of head_dim ** -0.5 if set.
        drop (float, optional): Dropout rate. Default: 0.0
        attn_drop (float, optional): Attention dropout rate. Default: 0.0
        drop_path (float | tuple[float], optional): Stochastic depth rate. Default: 0.0
        norm_layer (nn.Module, optional): Normalization layer. Default: nn.LayerNorm
        downsample (nn.Module | None, optional): Downsample layer at the end of the layer. Default: None
        use_checkpoint (bool): Whether to use checkpointing to save memory. Default: False.
    """

    def __init__(self, dim, input_resolution, depth, num_heads, window_size,
                 mlp_ratio=4., qkv_bias=True, qk_scale=None, drop=0., attn_drop=0.,
                 drop_path=0., norm_layer=nn.LayerNorm, downsample=None, use_checkpoint=False):

        super().__init__()
        self.dim = dim
        self.input_resolution = input_resolution
        self.depth = depth
        self.use_checkpoint = use_checkpoint

        # build blocks
        self.blocks = nn.ModuleList([
            SwinTransformerBlock(dim=dim, input_resolution=input_resolution,
                                 num_heads=num_heads, window_size=window_size,
                                 shift_size=0 if (i % 2 == 0) else window_size // 2,
                                 mlp_ratio=mlp_ratio,
                                 qkv_bias=qkv_bias, qk_scale=qk_scale,
                                 drop=drop, attn_drop=attn_drop,
                                 drop_path=drop_path[i] if isinstance(drop_path, list) else drop_path,
                                 norm_layer=norm_layer)
            for i in range(depth)])

        # patch merging layer
        if downsample is not None:
            self.downsample = downsample(input_resolution, dim=dim, norm_layer=norm_layer)
        else:
            self.downsample = None

    def forward(self, x):

        for blk in self.blocks:
            if self.use_checkpoint:
                x = checkpoint.checkpoint(blk, x)
            else:
                x = blk(x)
        if self.downsample is not None:
            x_down = self.downsample(x)
        elif self.downsample is None:
            x_down = x
        return x_down, x

    def extra_repr(self) -> str:
        return f"dim={self.dim}, input_resolution={self.input_resolution}, depth={self.depth}"

    def flops(self):
        flops = 0
        for blk in self.blocks:
            flops += blk.flops()
        if self.downsample is not None:
            flops += self.downsample.flops()
        return flops


class PatchEmbed(nn.Module):

    def __init__(self, img_size=224, patch_size=4, in_chans=3, embed_dim=96, norm_layer=None):
        super().__init__()
        img_size = to_2tuple(img_size)
        patch_size = to_2tuple(patch_size)
        patches_resolution = [img_size[0] // patch_size[0], img_size[1] // patch_size[1]]
        self.img_size = img_size
        self.patch_size = patch_size
        self.patches_resolution = patches_resolution
        self.num_patches = patches_resolution[0] * patches_resolution[1]

        self.in_chans = in_chans # define in_chans == 3
        self.embed_dim = embed_dim # Swin-B.embed_dim ==128,(T is 96)

        self.proj = nn.Conv2d(in_chans, embed_dim, kernel_size=patch_size, stride=patch_size) # dim 3->128
        if norm_layer is not None:
            self.norm = norm_layer(embed_dim)
        else:
            self.norm = None

    def forward(self, x):
        B, C, H, W = x.shape
        # FIXME look at relaxing size constraints,尺寸固定，下有断言
        assert H == self.img_size[0] and W == self.img_size[1], \
            f"Input image size ({H}*{W}) doesn't match model ({self.img_size[0]}*{self.img_size[1]})."

        x = self.proj(x).flatten(2).transpose(1, 2)  # B Ph*Pw C
        if self.norm is not None:
            x = self.norm(x)
        return x

    def flops(self):
        Ho, Wo = self.patches_resolution
        flops = Ho * Wo * self.embed_dim * self.in_chans * (self.patch_size[0] * self.patch_size[1])
        if self.norm is not None:
            flops += Ho * Wo * self.embed_dim
        return flops


class SwinTransformer(nn.Module):
    r""" Swin Transformer
        A PyTorch impl of : `Swin Transformer: Hierarchical Vision Transformer using Shifted Windows`  -
          https://arxiv.org/pdf/2103.14030

    Args:
        img_size (int | tuple(int)): Input image size. Default 224
        patch_size (int | tuple(int)): Patch size. Default: 4
        in_chans (int): Number of input image channels. Default: 3
        embed_dim (int): Patch embedding dimension. Default: 96
        depths (tuple(int)): Depth of each Swin Transformer layer.
        num_heads (tuple(int)): Number of attention heads in different layers.
        window_size (int): Window size. Default: 7
        mlp_ratio (float): Ratio of mlp hidden dim to embedding dim. Default: 4
        qkv_bias (bool): If True, add a learnable bias to query, key, value. Default: True
        qk_scale (float): Override default qk scale of head_dim ** -0.5 if set. Default: None
        drop_rate (float): Dropout rate. Default: 0
        attn_drop_rate (float): Attention dropout rate. Default: 0
        drop_path_rate (float): Stochastic depth rate. Default: 0.1
        norm_layer (nn.Module): Normalization layer. Default: nn.LayerNorm.
        ape (bool): If True, add absolute position embedding to the patch embedding. Default: False
        patch_norm (bool): If True, add normalization after patch embedding. Default: True
        use_checkpoint (bool): Whether to use checkpointing to save memory. Default: False
    """

    def __init__(self, img_size=384, patch_size=4, in_chans=3,
                 embed_dim=128, depths=[2, 2, 6, 2], num_heads=[3, 6, 12, 24],
                 window_size=12, mlp_ratio=4., qkv_bias=True, qk_scale=None,
                 drop_rate=0., attn_drop_rate=0., drop_path_rate=0.1,
                 norm_layer=nn.LayerNorm, ape=False, patch_norm=True,
                 use_checkpoint=False, **kwargs):
        super().__init__()


        self.num_layers = len(depths)
        self.embed_dim = embed_dim
        self.ape = ape
        self.patch_norm = patch_norm
        self.num_features = int(embed_dim * 2 ** (self.num_layers - 1))
        self.mlp_ratio = mlp_ratio


        self.patch_embed = PatchEmbed(
            img_size=img_size, patch_size=patch_size, in_chans=in_chans, embed_dim=embed_dim,
            norm_layer=norm_layer if self.patch_norm else None)
        num_patches = self.patch_embed.num_patches
        patches_resolution = self.patch_embed.patches_resolution
        self.patches_resolution = patches_resolution


        if self.ape:
            self.absolute_pos_embed = nn.Parameter(torch.zeros(1, num_patches, embed_dim))
            trunc_normal_(self.absolute_pos_embed, std=.02)

        self.pos_drop = nn.Dropout(p=drop_rate)


        dpr = [x.item() for x in torch.linspace(0, drop_path_rate, sum(depths))]  # stochastic depth decay rule

        # build layers
        self.layers = nn.ModuleList()
        for i_layer in range(self.num_layers):
            layer = BasicLayer(dim=int(embed_dim * 2 ** i_layer),
                               input_resolution=(patches_resolution[0] // (2 ** i_layer),
                                                 patches_resolution[1] // (2 ** i_layer)),
                               depth=depths[i_layer],
                               num_heads=num_heads[i_layer],
                               window_size=window_size,
                               mlp_ratio=self.mlp_ratio,
                               qkv_bias=qkv_bias, qk_scale=qk_scale,
                               drop=drop_rate, attn_drop=attn_drop_rate,
                               drop_path=dpr[sum(depths[:i_layer]):sum(depths[:i_layer + 1])],
                               norm_layer=norm_layer,
                               downsample=PatchMerging if (i_layer < self.num_layers - 1) else None,
                               use_checkpoint=use_checkpoint)
            # self.layers 中应该是 4 个
            self.layers.append(layer)

        self.norm = norm_layer(self.num_features)
        self.apply(self._init_weights)

    def _init_weights(self, m):
        if isinstance(m, nn.Linear):
            trunc_normal_(m.weight, std=.02)
            if isinstance(m, nn.Linear) and m.bias is not None:
                nn.init.constant_(m.bias, 0)
        elif isinstance(m, nn.LayerNorm):
            nn.init.constant_(m.bias, 0)
            nn.init.constant_(m.weight, 1.0)

    @torch.jit.ignore
    def no_weight_decay(self):
        return {'absolute_pos_embed'}

    @torch.jit.ignore
    def no_weight_decay_keywords(self):
        return {'relative_position_bias_table'}

    def forward_features(self, x):
        layer_features = []
        x = self.patch_embed(x)
        B,L,C = x.shape
        layer_features.append(x.view(B, int(np.sqrt(L)), int(np.sqrt(L)),-1).permute(0, 3, 1, 2).contiguous())
        # layer_features.append(x)
        if self.ape:
            x = x + self.absolute_pos_embed
        x = self.pos_drop(x)

        for layer in self.layers:
            x, x_undownsample = layer(x)
            B, L, C = x_undownsample.shape
            # print('x:', x.shape)
            xl = x_undownsample.view(B, int(np.sqrt(L)), int(np.sqrt(L)),-1).permute(0, 3, 1, 2).contiguous()
            # print('xl',xl.shape)
            layer_features.append(xl)
        x = self.norm(x)  # B L C
        B, L, C = x.shape
        # x = self.avgpool(x.ranspose(1, 2))  # B C 1
        x = x.view(B, int(np.sqrt(L)), int(np.sqrt(L)),-1).permute(0, 3, 1, 2).contiguous()
        # x = torch.flatten(x, 1)
        layer_features[-1] = x



        return layer_features

    def forward(self, x):
        outs = self.forward_features(x)

        return outs


class SwinMCNet(nn.Module):
    def __init__(self, norm_layer = nn.LayerNorm):
        super(SwinMCNet, self).__init__()

        self.swin_image = SwinTransformer(embed_dim=128, depths=[2,2,18,2], num_heads=[4,8,16,32])
        self.swin_thermal = SwinTransformer(embed_dim=128, depths=[2,2,18,2], num_heads=[4,8,16,32])
        
        self.bi5   = nn.Sequential(nn.Conv2d(1024, 64, kernel_size=1), nn.Conv2d(64, 64, kernel_size=3, padding=1), nn.BatchNorm2d(64), nn.ReLU(inplace=True))
        self.bi4   = nn.Sequential(nn.Conv2d( 512, 64, kernel_size=1), nn.Conv2d(64, 64, kernel_size=3, padding=1), nn.BatchNorm2d(64), nn.ReLU(inplace=True))
        self.bi3   = nn.Sequential(nn.Conv2d( 256, 64, kernel_size=1), nn.Conv2d(64, 64, kernel_size=3, padding=1), nn.BatchNorm2d(64), nn.ReLU(inplace=True))
        self.bi2   = nn.Sequential(nn.Conv2d( 128, 64, kernel_size=1), nn.Conv2d(64, 64, kernel_size=3, padding=1), nn.BatchNorm2d(64), nn.ReLU(inplace=True))
        
        self.bt5   = nn.Sequential(nn.Conv2d(1024, 64, kernel_size=1), nn.Conv2d(64, 64, kernel_size=3, padding=1), nn.BatchNorm2d(64), nn.ReLU(inplace=True))
        self.bt4   = nn.Sequential(nn.Conv2d( 512, 64, kernel_size=1), nn.Conv2d(64, 64, kernel_size=3, padding=1), nn.BatchNorm2d(64), nn.ReLU(inplace=True))
        self.bt3   = nn.Sequential(nn.Conv2d( 256, 64, kernel_size=1), nn.Conv2d(64, 64, kernel_size=3, padding=1), nn.BatchNorm2d(64), nn.ReLU(inplace=True))
        self.bt2   = nn.Sequential(nn.Conv2d( 128, 64, kernel_size=1), nn.Conv2d(64, 64, kernel_size=3, padding=1), nn.BatchNorm2d(64), nn.ReLU(inplace=True))

        self.att_2 = Attention()
        self.att_3 = Attention()
        self.att_4 = Attention()
        self.att_5 = Attention()

        #self.encoder = Encoder()
        self.rfb = LightRFB()
        self.decoderi = Decoder()
        self.decodert = Decoder()
        self.lineari  = nn.Conv2d(64, 1, kernel_size=3, padding=1)
        self.lineart  = nn.Conv2d(64, 1, kernel_size=3, padding=1)
        self.linear   = nn.Sequential(nn.Conv2d(128, 64, kernel_size=3, padding=1), nn.BatchNorm2d(64), nn.ReLU(inplace=True), nn.Conv2d(64, 1, kernel_size=3, padding=1))

    def forward(self, image, thermal, shape=None):
        image_list = self.swin_image(image)
        thermal_list = self.swin_thermal(thermal)
        
        i1,i2,i3,i4,i5 = image_list[0], image_list[1], image_list[2], image_list[3], image_list[4]
        t1,t2,t3,t4,t5 = thermal_list[0], thermal_list[1], thermal_list[2], thermal_list[3], thermal_list[4]

        i2,i3,i4,i5 = self.bi2(i2), self.bi3(i3), self.bi4(i4), self.bi5(i5)
        t2,t3,t4,t5 = self.bt2(t2), self.bt3(t3), self.bt4(t4), self.bt5(t5)

        
        att2i, att2t = self.att_2(i2, t2)
        att3i, att3t = self.att_3(i3, t3)
        att4i, att4t = self.att_4(i4, t4)
        att5i, att5t = self.att_5(i5, t5)

        out2i, out3i, out4i, out5i = i2 + att2i, i3 + att3i, i4 + att4i, i5 + att5i
        out2t, out3t, out4t, out5t = t2 + att2t, t3 + att3t, t4 + att4t, t5 + att5t

        outi1 = self.decoderi([out5i, out4i, out3i, out2i])
        outt1 = self.decodert([out5t, out4t, out3t, out2t])

        out1  = torch.cat([outi1, outt1], dim=1)

        outi2, outt2 = self.rfb(out1)

        outi2 = self.decoderi([out5i, out4i, out3i, out2i], outi2)
        outt2 = self.decodert([out5t, out4t, out3t, out2t], outt2)

        out2  = torch.cat([outi2, outt2], dim=1)

        if shape is None:
            shape = image.size()[2:]
        out1  = F.interpolate(self.linear(out1),   size=shape, mode='bilinear')
        outi1 = F.interpolate(self.lineari(outi1), size=shape, mode='bilinear')
        outt1 = F.interpolate(self.lineart(outt1), size=shape, mode='bilinear')

        out2  = F.interpolate(self.linear(out2),   size=shape, mode='bilinear')
        outi2 = F.interpolate(self.lineari(outi2), size=shape, mode='bilinear')
        outt2 = F.interpolate(self.lineart(outt2), size=shape, mode='bilinear')
        
        #return outi1, outt1, out1, outi2, outt2, out2
        return i2,i3,i4,i5,t2,t3,t4,t5

    def load_pre(self, pre_model):
        self.swin_image.load_state_dict(torch.load(pre_model)['model'],strict=False)
        print(f"RGB SwinTransformer loading pre_model ${pre_model}")
        self.swin_thermal.load_state_dict(torch.load(pre_model)['model'], strict=False)
        print(f"Thermal SwinTransformer loading pre_model ${pre_model}")