YOLO11改进|注意力机制篇|引入HAT超分辨率重建模块

一、HAttention注意力机制

1.1HAttention注意力介绍

HAT模型 通过结合卷积特征提取与多尺度注意力机制，具备了强大的图像重建能力。它的优势在于能有效整合局部和全局信息，并通过残差连接和通道注意力等方式提高网络的表达能力和重建质量，适用于图像超分辨率和图像重建任务。
下面是HAT的工作流程和主要模块的作用

• 浅层特征提取 (Shallow Feature Extraction)
  输入图像首先经过卷积操作提取低级特征。该过程用来捕捉图像的基础信息，如边缘、颜色等，形成初步的特征图。
• 深层特征提取 (Deep Feature Extraction)
  浅层特征通过多个RHAG模块进行深度特征提取。RHAG由多个HAB（混合注意力块）和OCAB（重叠交叉注意力块）组成：
  HAB：包含 CAB (Channel Attention Block) 和 (S)W-MSA (Shifted Window Multi-Head Self-Attention) 结构。
  CAB (通道注意力块) 使用全局池化和通道注意力机制，专注于不同通道之间的依赖关系，以增强特定通道的特征表示。
  (S)W-MSA 是一种窗口划分的自注意力机制，通过窗口化操作计算注意力，减少计算开销，同时增强局部与全局信息的交互。
  OCAB：通过交叉注意力机制结合局部和全局特征，并通过重叠区域确保信息的连贯性和连续性。
  优势：深度特征提取模块通过多个注意力模块结合局部和全局信息，实现对复杂特征的高效捕捉，同时保持较低的计算成本。
• 图像重建 (Image Reconstruction)
  深层特征经过多个RHAG模块后，通过上采样操作重建回高分辨率图像。模型将提取到的深层特征与初始输入进行特征融合，生成更高质量的重建图像。
• 模块优势
  RHAG (Residual Hybrid Attention Group)：该模块通过残差连接增强网络的梯度流，避免深层网络中的梯度消失问题，同时结合多种注意力机制，提高特征提取的准确性和效率。
  HAB (Hybrid Attention Block)：该模块将通道注意力与窗口自注意力相结合，在不同尺度上捕捉图像特征。通道注意力增强了各个特征通道的表示能力，而窗口自注意力通过局部和全局上下文的信息交互来提升整体的特征感知能力。
  OCAB (Overlapping Cross-Attention Block)：通过交叉注意力和重叠区域融合，使模型在捕捉局部特征的同时，能够保持对全局特征的感知，避免信息的割裂。

1.2HAT核心代码

import math
import torch
import torch.nn as nn
from basicsr.utils.registry import ARCH_REGISTRY
from basicsr.archs.arch_util import to_2tuple, trunc_normal_
from einops import rearrangedef drop_path(x, drop_prob: float = 0., training: bool = False):"""Drop paths (Stochastic Depth) per sample (when applied in main path of residual blocks).From: https://github.com/rwightman/pytorch-image-models/blob/master/timm/models/layers/drop.py"""if drop_prob == 0. or not training:return xkeep_prob = 1 - drop_probshape = (x.shape[0], ) + (1, ) * (x.ndim - 1)  # work with diff dim tensors, not just 2D ConvNetsrandom_tensor = keep_prob + torch.rand(shape, dtype=x.dtype, device=x.device)random_tensor.floor_()  # binarizeoutput = x.div(keep_prob) * random_tensorreturn outputclass DropPath(nn.Module):"""Drop paths (Stochastic Depth) per sample  (when applied in main path of residual blocks).From: https://github.com/rwightman/pytorch-image-models/blob/master/timm/models/layers/drop.py"""def __init__(self, drop_prob=None):super(DropPath, self).__init__()self.drop_prob = drop_probdef forward(self, x):return drop_path(x, self.drop_prob, self.training)class ChannelAttention(nn.Module):"""Channel attention used in RCAN.Args:num_feat (int): Channel number of intermediate features.squeeze_factor (int): Channel squeeze factor. Default: 16."""def __init__(self, num_feat, squeeze_factor=16):super(ChannelAttention, self).__init__()self.attention = nn.Sequential(nn.AdaptiveAvgPool2d(1),nn.Conv2d(num_feat, num_feat // squeeze_factor, 1, padding=0),nn.ReLU(inplace=True),nn.Conv2d(num_feat // squeeze_factor, num_feat, 1, padding=0),nn.Sigmoid())def forward(self, x):y = self.attention(x)return x * yclass CAB(nn.Module):def __init__(self, num_feat, compress_ratio=3, squeeze_factor=30):super(CAB, self).__init__()self.cab = nn.Sequential(nn.Conv2d(num_feat, num_feat // compress_ratio, 3, 1, 1),nn.GELU(),nn.Conv2d(num_feat // compress_ratio, num_feat, 3, 1, 1),ChannelAttention(num_feat, squeeze_factor))def forward(self, x):return self.cab(x)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_featureshidden_features = hidden_features or in_featuresself.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 xdef window_partition(x, window_size):"""Args:x: (b, h, w, c)window_size (int): window sizeReturns:windows: (num_windows*b, window_size, window_size, c)"""b, h, w, c = x.shapex = 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 windowsdef window_reverse(windows, window_size, h, w):"""Args:windows: (num_windows*b, window_size, window_size, c)window_size (int): Window sizeh (int): Height of imagew (int): Width of imageReturns: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 xclass 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: Trueqk_scale (float | None, optional): Override default qk scale of head_dim ** -0.5 if setattn_drop (float, optional): Dropout ratio of attention weight. Default: 0.0proj_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 = dimself.window_size = window_size  # Wh, Wwself.num_heads = num_headshead_dim = dim // num_headsself.scale = qk_scale or head_dim**-0.5# define a parameter table of relative position biasself.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, nHself.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, rpi, 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.shapeqkv = 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.scaleattn = (q @ k.transpose(-2, -1))relative_position_bias = self.relative_position_bias_table[rpi.view(-1)].view(self.window_size[0] * self.window_size[1], self.window_size[0] * self.window_size[1], -1)  # Wh*Ww,Wh*Ww,nHrelative_position_bias = relative_position_bias.permute(2, 0, 1).contiguous()  # nH, Wh*Ww, Wh*Wwattn = 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 xclass HAB(nn.Module):r""" Hybrid Attention Block.Args:dim (int): Number of input channels.input_resolution (tuple[int]): Input resolution.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: Trueqk_scale (float | None, optional): Override default qk scale of head_dim ** -0.5 if set.drop (float, optional): Dropout rate. Default: 0.0attn_drop (float, optional): Attention dropout rate. Default: 0.0drop_path (float, optional): Stochastic depth rate. Default: 0.0act_layer (nn.Module, optional): Activation layer. Default: nn.GELUnorm_layer (nn.Module, optional): Normalization layer.  Default: nn.LayerNorm"""def __init__(self,dim,input_resolution,num_heads,window_size=7,shift_size=0,compress_ratio=3,squeeze_factor=30,conv_scale=0.01,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 = dimself.input_resolution = input_resolutionself.num_heads = num_headsself.window_size = window_sizeself.shift_size = shift_sizeself.mlp_ratio = mlp_ratioif min(self.input_resolution) <= self.window_size:# if window size is larger than input resolution, we don't partition windowsself.shift_size = 0self.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.conv_scale = conv_scaleself.conv_block = CAB(num_feat=dim, compress_ratio=compress_ratio, squeeze_factor=squeeze_factor)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)def forward(self, x, x_size, rpi_sa, attn_mask):h, w = x_sizeb, _, c = x.shape# assert seq_len == h * w, "input feature has wrong size"shortcut = xx = self.norm1(x)x = x.view(b, h, w, c)# Conv_Xconv_x = self.conv_block(x.permute(0, 3, 1, 2))conv_x = conv_x.permute(0, 2, 3, 1).contiguous().view(b, h * w, c)# cyclic shiftif self.shift_size > 0:shifted_x = torch.roll(x, shifts=(-self.shift_size, -self.shift_size), dims=(1, 2))attn_mask = attn_maskelse:shifted_x = xattn_mask = None# partition windowsx_windows = window_partition(shifted_x, self.window_size)  # nw*b, window_size, window_size, cx_windows = x_windows.view(-1, self.window_size * self.window_size, c)  # nw*b, window_size*window_size, c# W-MSA/SW-MSA (to be compatible for testing on images whose shapes are the multiple of window sizeattn_windows = self.attn(x_windows, rpi=rpi_sa, mask=attn_mask)# merge windowsattn_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 shiftif self.shift_size > 0:attn_x = torch.roll(shifted_x, shifts=(self.shift_size, self.shift_size), dims=(1, 2))else:attn_x = shifted_xattn_x = attn_x.view(b, h * w, c)# FFNx = shortcut + self.drop_path(attn_x) + conv_x * self.conv_scalex = x + self.drop_path(self.mlp(self.norm2(x)))return xclass 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_resolutionself.dim = dimself.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_resolutionb, seq_len, c = x.shapeassert seq_len == 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 cx1 = x[:, 1::2, 0::2, :]  # b h/2 w/2 cx2 = x[:, 0::2, 1::2, :]  # b h/2 w/2 cx3 = x[:, 1::2, 1::2, :]  # b h/2 w/2 cx = torch.cat([x0, x1, x2, x3], -1)  # b h/2 w/2 4*cx = x.view(b, -1, 4 * c)  # b h/2*w/2 4*cx = self.norm(x)x = self.reduction(x)return xclass OCAB(nn.Module):# overlapping cross-attention blockdef __init__(self, dim,input_resolution,window_size,overlap_ratio,num_heads,qkv_bias=True,qk_scale=None,mlp_ratio=2,norm_layer=nn.LayerNorm):super().__init__()self.dim = dimself.input_resolution = input_resolutionself.window_size = window_sizeself.num_heads = num_headshead_dim = dim // num_headsself.scale = qk_scale or head_dim**-0.5self.overlap_win_size = int(window_size * overlap_ratio) + window_sizeself.norm1 = norm_layer(dim)self.qkv = nn.Linear(dim, dim * 3,  bias=qkv_bias)self.unfold = nn.Unfold(kernel_size=(self.overlap_win_size, self.overlap_win_size), stride=window_size, padding=(self.overlap_win_size-window_size)//2)# define a parameter table of relative position biasself.relative_position_bias_table = nn.Parameter(torch.zeros((window_size + self.overlap_win_size - 1) * (window_size + self.overlap_win_size - 1), num_heads))  # 2*Wh-1 * 2*Ww-1, nHtrunc_normal_(self.relative_position_bias_table, std=.02)self.softmax = nn.Softmax(dim=-1)self.proj = nn.Linear(dim,dim)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=nn.GELU)def forward(self, x, x_size, rpi):h, w = x_sizeb, _, c = x.shapeshortcut = xx = self.norm1(x)x = x.view(b, h, w, c)qkv = self.qkv(x).reshape(b, h, w, 3, c).permute(3, 0, 4, 1, 2) # 3, b, c, h, wq = qkv[0].permute(0, 2, 3, 1) # b, h, w, ckv = torch.cat((qkv[1], qkv[2]), dim=1) # b, 2*c, h, w# partition windowsq_windows = window_partition(q, self.window_size)  # nw*b, window_size, window_size, cq_windows = q_windows.view(-1, self.window_size * self.window_size, c)  # nw*b, window_size*window_size, ckv_windows = self.unfold(kv) # b, c*w*w, nwkv_windows = rearrange(kv_windows, 'b (nc ch owh oww) nw -> nc (b nw) (owh oww) ch', nc=2, ch=c, owh=self.overlap_win_size, oww=self.overlap_win_size).contiguous() # 2, nw*b, ow*ow, ck_windows, v_windows = kv_windows[0], kv_windows[1] # nw*b, ow*ow, cb_, nq, _ = q_windows.shape_, n, _ = k_windows.shaped = self.dim // self.num_headsq = q_windows.reshape(b_, nq, self.num_heads, d).permute(0, 2, 1, 3) # nw*b, nH, nq, dk = k_windows.reshape(b_, n, self.num_heads, d).permute(0, 2, 1, 3) # nw*b, nH, n, dv = v_windows.reshape(b_, n, self.num_heads, d).permute(0, 2, 1, 3) # nw*b, nH, n, dq = q * self.scaleattn = (q @ k.transpose(-2, -1))relative_position_bias = self.relative_position_bias_table[rpi.view(-1)].view(self.window_size * self.window_size, self.overlap_win_size * self.overlap_win_size, -1)  # ws*ws, wse*wse, nHrelative_position_bias = relative_position_bias.permute(2, 0, 1).contiguous()  # nH, ws*ws, wse*wseattn = attn + relative_position_bias.unsqueeze(0)attn = self.softmax(attn)attn_windows = (attn @ v).transpose(1, 2).reshape(b_, nq, self.dim)# merge windowsattn_windows = attn_windows.view(-1, self.window_size, self.window_size, self.dim)x = window_reverse(attn_windows, self.window_size, h, w)  # b h w cx = x.view(b, h * w, self.dim)x = self.proj(x) + shortcutx = x + self.mlp(self.norm2(x))return xclass AttenBlocks(nn.Module):""" A series of attention blocks for one RHAG.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: Trueqk_scale (float | None, optional): Override default qk scale of head_dim ** -0.5 if set.drop (float, optional): Dropout rate. Default: 0.0attn_drop (float, optional): Attention dropout rate. Default: 0.0drop_path (float | tuple[float], optional): Stochastic depth rate. Default: 0.0norm_layer (nn.Module, optional): Normalization layer. Default: nn.LayerNormdownsample (nn.Module | None, optional): Downsample layer at the end of the layer. Default: Noneuse_checkpoint (bool): Whether to use checkpointing to save memory. Default: False."""def __init__(self,dim,input_resolution,depth,num_heads,window_size,compress_ratio,squeeze_factor,conv_scale,overlap_ratio,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 = dimself.input_resolution = input_resolutionself.depth = depthself.use_checkpoint = use_checkpoint# build blocksself.blocks = nn.ModuleList([HAB(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,compress_ratio=compress_ratio,squeeze_factor=squeeze_factor,conv_scale=conv_scale,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)])# OCABself.overlap_attn = OCAB(dim=dim,input_resolution=input_resolution,window_size=window_size,overlap_ratio=overlap_ratio,num_heads=num_heads,qkv_bias=qkv_bias,qk_scale=qk_scale,mlp_ratio=mlp_ratio,norm_layer=norm_layer)# patch merging layerif downsample is not None:self.downsample = downsample(input_resolution, dim=dim, norm_layer=norm_layer)else:self.downsample = Nonedef forward(self, x, x_size, params):for blk in self.blocks:x = blk(x, x_size, params['rpi_sa'], params['attn_mask'])x = self.overlap_attn(x, x_size, params['rpi_oca'])if self.downsample is not None:x = self.downsample(x)return xclass RHAG(nn.Module):"""Residual Hybrid Attention Group (RHAG).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: Trueqk_scale (float | None, optional): Override default qk scale of head_dim ** -0.5 if set.drop (float, optional): Dropout rate. Default: 0.0attn_drop (float, optional): Attention dropout rate. Default: 0.0drop_path (float | tuple[float], optional): Stochastic depth rate. Default: 0.0norm_layer (nn.Module, optional): Normalization layer. Default: nn.LayerNormdownsample (nn.Module | None, optional): Downsample layer at the end of the layer. Default: Noneuse_checkpoint (bool): Whether to use checkpointing to save memory. Default: False.img_size: Input image size.patch_size: Patch size.resi_connection: The convolutional block before residual connection."""def __init__(self,dim,input_resolution,depth,num_heads,window_size,compress_ratio,squeeze_factor,conv_scale,overlap_ratio,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,img_size=224,patch_size=4,resi_connection='1conv'):super(RHAG, self).__init__()self.dim = dimself.input_resolution = input_resolutionself.residual_group = AttenBlocks(dim=dim,input_resolution=input_resolution,depth=depth,num_heads=num_heads,window_size=window_size,compress_ratio=compress_ratio,squeeze_factor=squeeze_factor,conv_scale=conv_scale,overlap_ratio=overlap_ratio,mlp_ratio=mlp_ratio,qkv_bias=qkv_bias,qk_scale=qk_scale,drop=drop,attn_drop=attn_drop,drop_path=drop_path,norm_layer=norm_layer,downsample=downsample,use_checkpoint=use_checkpoint)if resi_connection == '1conv':self.conv = nn.Conv2d(dim, dim, 3, 1, 1)elif resi_connection == 'identity':self.conv = nn.Identity()self.patch_embed = PatchEmbed(img_size=img_size, patch_size=patch_size, in_chans=0, embed_dim=dim, norm_layer=None)self.patch_unembed = PatchUnEmbed(img_size=img_size, patch_size=patch_size, in_chans=0, embed_dim=dim, norm_layer=None)def forward(self, x, x_size, params):return self.patch_embed(self.conv(self.patch_unembed(self.residual_group(x, x_size, params), x_size))) + xclass PatchEmbed(nn.Module):r""" Image to Patch EmbeddingArgs:img_size (int): Image size.  Default: 224.patch_size (int): Patch token size. Default: 4.in_chans (int): Number of input image channels. Default: 3.embed_dim (int): Number of linear projection output channels. Default: 96.norm_layer (nn.Module, optional): Normalization layer. Default: None"""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_sizeself.patch_size = patch_sizeself.patches_resolution = patches_resolutionself.num_patches = patches_resolution[0] * patches_resolution[1]self.in_chans = in_chansself.embed_dim = embed_dimif norm_layer is not None:self.norm = norm_layer(embed_dim)else:self.norm = Nonedef forward(self, x):x = x.flatten(2).transpose(1, 2)  # b Ph*Pw cif self.norm is not None:x = self.norm(x)return xclass PatchUnEmbed(nn.Module):r""" Image to Patch UnembeddingArgs:img_size (int): Image size.  Default: 224.patch_size (int): Patch token size. Default: 4.in_chans (int): Number of input image channels. Default: 3.embed_dim (int): Number of linear projection output channels. Default: 96.norm_layer (nn.Module, optional): Normalization layer. Default: None"""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_sizeself.patch_size = patch_sizeself.patches_resolution = patches_resolutionself.num_patches = patches_resolution[0] * patches_resolution[1]self.in_chans = in_chansself.embed_dim = embed_dimdef forward(self, x, x_size):x = x.transpose(1, 2).contiguous().view(x.shape[0], self.embed_dim, x_size[0], x_size[1])  # b Ph*Pw creturn xclass Upsample(nn.Sequential):"""Upsample module.Args:scale (int): Scale factor. Supported scales: 2^n and 3.num_feat (int): Channel number of intermediate features."""def __init__(self, scale, num_feat):m = []if (scale & (scale - 1)) == 0:  # scale = 2^nfor _ in range(int(math.log(scale, 2))):m.append(nn.Conv2d(num_feat, 4 * num_feat, 3, 1, 1))m.append(nn.PixelShuffle(2))elif scale == 3:m.append(nn.Conv2d(num_feat, 9 * num_feat, 3, 1, 1))m.append(nn.PixelShuffle(3))else:raise ValueError(f'scale {scale} is not supported. ' 'Supported scales: 2^n and 3.')super(Upsample, self).__init__(*m)@ARCH_REGISTRY.register()
class HAT(nn.Module):r""" Hybrid Attention TransformerA PyTorch implementation of : `Activating More Pixels in Image Super-Resolution Transformer`.Some codes are based on SwinIR.Args:img_size (int | tuple(int)): Input image size. Default 64patch_size (int | tuple(int)): Patch size. Default: 1in_chans (int): Number of input image channels. Default: 3embed_dim (int): Patch embedding dimension. Default: 96depths (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: 7mlp_ratio (float): Ratio of mlp hidden dim to embedding dim. Default: 4qkv_bias (bool): If True, add a learnable bias to query, key, value. Default: Trueqk_scale (float): Override default qk scale of head_dim ** -0.5 if set. Default: Nonedrop_rate (float): Dropout rate. Default: 0attn_drop_rate (float): Attention dropout rate. Default: 0drop_path_rate (float): Stochastic depth rate. Default: 0.1norm_layer (nn.Module): Normalization layer. Default: nn.LayerNorm.ape (bool): If True, add absolute position embedding to the patch embedding. Default: Falsepatch_norm (bool): If True, add normalization after patch embedding. Default: Trueuse_checkpoint (bool): Whether to use checkpointing to save memory. Default: Falseupscale: Upscale factor. 2/3/4/8 for image SR, 1 for denoising and compress artifact reductionimg_range: Image range. 1. or 255.upsampler: The reconstruction reconstruction module. 'pixelshuffle'/'pixelshuffledirect'/'nearest+conv'/Noneresi_connection: The convolutional block before residual connection. '1conv'/'3conv'"""def __init__(self,in_chans=3,img_size=64,patch_size=1,embed_dim=96,depths=(6, 6, 6, 6),num_heads=(6, 6, 6, 6),window_size=7,compress_ratio=3,squeeze_factor=30,conv_scale=0.01,overlap_ratio=0.5,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,upscale=2,img_range=1.,upsampler='',resi_connection='1conv',**kwargs):super(HAT, self).__init__()self.window_size = window_sizeself.shift_size = window_size // 2self.overlap_ratio = overlap_rationum_in_ch = in_chansnum_out_ch = in_chansnum_feat = 64self.img_range = img_rangeif in_chans == 3:rgb_mean = (0.4488, 0.4371, 0.4040)self.mean = torch.Tensor(rgb_mean).view(1, 3, 1, 1)else:self.mean = torch.zeros(1, 1, 1, 1)self.upscale = upscaleself.upsampler = upsampler# relative position indexrelative_position_index_SA = self.calculate_rpi_sa()relative_position_index_OCA = self.calculate_rpi_oca()self.register_buffer('relative_position_index_SA', relative_position_index_SA)self.register_buffer('relative_position_index_OCA', relative_position_index_OCA)# ------------------------- 1, shallow feature extraction ------------------------- #self.conv_first = nn.Conv2d(num_in_ch, embed_dim, 3, 1, 1)# ------------------------- 2, deep feature extraction ------------------------- #self.num_layers = len(depths)self.embed_dim = embed_dimself.ape = apeself.patch_norm = patch_normself.num_features = embed_dimself.mlp_ratio = mlp_ratio# split image into non-overlapping patchesself.patch_embed = PatchEmbed(img_size=img_size,patch_size=patch_size,in_chans=embed_dim,embed_dim=embed_dim,norm_layer=norm_layer if self.patch_norm else None)num_patches = self.patch_embed.num_patchespatches_resolution = self.patch_embed.patches_resolutionself.patches_resolution = patches_resolution# merge non-overlapping patches into imageself.patch_unembed = PatchUnEmbed(img_size=img_size,patch_size=patch_size,in_chans=embed_dim,embed_dim=embed_dim,norm_layer=norm_layer if self.patch_norm else None)# absolute position embeddingif 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)# stochastic depthdpr = [x.item() for x in torch.linspace(0, drop_path_rate, sum(depths))]  # stochastic depth decay rule# build Residual Hybrid Attention Groups (RHAG)self.layers = nn.ModuleList()for i_layer in range(self.num_layers):layer = RHAG(dim=embed_dim,input_resolution=(patches_resolution[0], patches_resolution[1]),depth=depths[i_layer],num_heads=num_heads[i_layer],window_size=window_size,compress_ratio=compress_ratio,squeeze_factor=squeeze_factor,conv_scale=conv_scale,overlap_ratio=overlap_ratio,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])],  # no impact on SR resultsnorm_layer=norm_layer,downsample=None,use_checkpoint=use_checkpoint,img_size=img_size,patch_size=patch_size,resi_connection=resi_connection)self.layers.append(layer)self.norm = norm_layer(self.num_features)# build the last conv layer in deep feature extractionif resi_connection == '1conv':self.conv_after_body = nn.Conv2d(embed_dim, embed_dim, 3, 1, 1)elif resi_connection == 'identity':self.conv_after_body = nn.Identity()# ------------------------- 3, high quality image reconstruction ------------------------- #if self.upsampler == 'pixelshuffle':# for classical SRself.conv_before_upsample = nn.Sequential(nn.Conv2d(embed_dim, num_feat, 3, 1, 1), nn.LeakyReLU(inplace=True))self.upsample = Upsample(upscale, num_feat)self.conv_last = nn.Conv2d(num_feat, num_out_ch, 3, 1, 1)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)def calculate_rpi_sa(self):# calculate relative position index for SAcoords_h = torch.arange(self.window_size)coords_w = torch.arange(self.window_size)coords = torch.stack(torch.meshgrid([coords_h, coords_w]))  # 2, Wh, Wwcoords_flatten = torch.flatten(coords, 1)  # 2, Wh*Wwrelative_coords = coords_flatten[:, :, None] - coords_flatten[:, None, :]  # 2, Wh*Ww, Wh*Wwrelative_coords = relative_coords.permute(1, 2, 0).contiguous()  # Wh*Ww, Wh*Ww, 2relative_coords[:, :, 0] += self.window_size - 1  # shift to start from 0relative_coords[:, :, 1] += self.window_size - 1relative_coords[:, :, 0] *= 2 * self.window_size - 1relative_position_index = relative_coords.sum(-1)  # Wh*Ww, Wh*Wwreturn relative_position_indexdef calculate_rpi_oca(self):# calculate relative position index for OCAwindow_size_ori = self.window_sizewindow_size_ext = self.window_size + int(self.overlap_ratio * self.window_size)coords_h = torch.arange(window_size_ori)coords_w = torch.arange(window_size_ori)coords_ori = torch.stack(torch.meshgrid([coords_h, coords_w]))  # 2, ws, wscoords_ori_flatten = torch.flatten(coords_ori, 1)  # 2, ws*wscoords_h = torch.arange(window_size_ext)coords_w = torch.arange(window_size_ext)coords_ext = torch.stack(torch.meshgrid([coords_h, coords_w]))  # 2, wse, wsecoords_ext_flatten = torch.flatten(coords_ext, 1)  # 2, wse*wserelative_coords = coords_ext_flatten[:, None, :] - coords_ori_flatten[:, :, None]   # 2, ws*ws, wse*wserelative_coords = relative_coords.permute(1, 2, 0).contiguous()  # ws*ws, wse*wse, 2relative_coords[:, :, 0] += window_size_ori - window_size_ext + 1  # shift to start from 0relative_coords[:, :, 1] += window_size_ori - window_size_ext + 1relative_coords[:, :, 0] *= window_size_ori + window_size_ext - 1relative_position_index = relative_coords.sum(-1)return relative_position_indexdef calculate_mask(self, x_size):# calculate attention mask for SW-MSAh, w = x_sizeimg_mask = torch.zeros((1, h, w, 1))  # 1 h w 1h_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 = 0for h in h_slices:for w in w_slices:img_mask[:, h, w, :] = cntcnt += 1mask_windows = window_partition(img_mask, self.window_size)  # nw, window_size, window_size, 1mask_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))return [e-mail] no_weight_decay(self):return {'absolute_pos_embed'}@torch.jit.ignoredef no_weight_decay_keywords(self):return {'relative_position_bias_table'}def forward_features(self, x):x_size = (x.shape[2], x.shape[3])# Calculate attention mask and relative position index in advance to speed up inference.# The original code is very time-consuming for large window size.attn_mask = self.calculate_mask(x_size).to(x.device)params = {'attn_mask': attn_mask, 'rpi_sa': self.relative_position_index_SA, 'rpi_oca': self.relative_position_index_OCA}x = self.patch_embed(x)if self.ape:x = x + self.absolute_pos_embedx = self.pos_drop(x)for layer in self.layers:x = layer(x, x_size, params)x = self.norm(x)  # b seq_len cx = self.patch_unembed(x, x_size)return xdef forward(self, x):self.mean = self.mean.type_as(x)x = (x - self.mean) * self.img_rangeif self.upsampler == 'pixelshuffle':# for classical SRx = self.conv_first(x)x = self.conv_after_body(self.forward_features(x)) + xx = self.conv_before_upsample(x)x = self.conv_last(self.upsample(x))x = x / self.img_range + self.meanreturn x

二、添加HAT注意力机制

2.1STEP1

首先找到ultralytics/nn文件路径下新建一个Add-module的python文件包【这里注意一定是python文件包，新建后会自动生成_init_.py】，如果已经跟着我的教程建立过一次了可以省略此步骤，随后新建一个HAT.py文件并将上文中提到的注意力机制的代码全部粘贴到此文件中，如下图所示

2.2STEP2

在STEP1中新建的_init_.py文件中导入增加改进模块的代码包如下图所示

2.3STEP3

找到ultralytics/nn文件夹中的task.py文件，在其中按照下图添加

2.4STEP4

定位到ultralytics/nn文件夹中的task.py文件中的def parse_model(d, ch, verbose=True): # model_dict, input_channels(3)函数添加如图代码,【如果不好定位可以直接ctrl+f搜索定位】

三、yaml文件与运行

3.1yaml文件

以下是添加HAT注意力机制在Backbone中的yaml文件，大家可以注释自行调节，效果以自己的数据集结果为准

# Ultralytics YOLO 🚀, AGPL-3.0 license
# YOLO11 object detection model with P3-P5 outputs. For Usage examples see https://docs.ultralytics.com/tasks/detect# Parameters
nc: 80 # number of classes
scales: # model compound scaling constants, i.e. 'model=yolo11n.yaml' will call yolo11.yaml with scale 'n'# [depth, width, max_channels]n: [0.50, 0.25, 1024] # summary: 319 layers, 2624080 parameters, 2624064 gradients, 6.6 GFLOPss: [0.50, 0.50, 1024] # summary: 319 layers, 9458752 parameters, 9458736 gradients, 21.7 GFLOPsm: [0.50, 1.00, 512] # summary: 409 layers, 20114688 parameters, 20114672 gradients, 68.5 GFLOPsl: [1.00, 1.00, 512] # summary: 631 layers, 25372160 parameters, 25372144 gradients, 87.6 GFLOPsx: [1.00, 1.50, 512] # summary: 631 layers, 56966176 parameters, 56966160 gradients, 196.0 GFLOPs# YOLO11n backbone
backbone:# [from, repeats, module, args]- [-1, 1, Conv, [64, 3, 2]] # 0-P1/2- [-1, 1, Conv, [128, 3, 2]] # 1-P2/4- [-1, 2, C3k2, [256, False, 0.25]]- [-1, 1, Conv, [256, 3, 2]] # 3-P3/8- [-1, 2, C3k2, [512, False, 0.25]]- [-1, 1, Conv, [512, 3, 2]] # 5-P4/16- [-1, 2, C3k2, [512, True]]- [-1, 1, Conv, [1024, 3, 2]] # 7-P5/32- [-1, 2, C3k2, [1024, True]]- [-1, 1, HAT,  []]- [-1, 1, SPPF, [1024, 5]] # 9- [-1, 2, C2PSA, [1024]] # 10# YOLO11n head
head:- [-1, 1, nn.Upsample, [None, 2, "nearest"]]- [[-1, 6], 1, Concat, [1]] # cat backbone P4- [-1, 2, C3k2, [512, False]] # 13- [-1, 1, nn.Upsample, [None, 2, "nearest"]]- [[-1, 4], 1, Concat, [1]] # cat backbone P3- [-1, 2, C3k2, [256, False]] # 16 (P3/8-small)- [-1, 1, Conv, [256, 3, 2]]- [[-1, 14], 1, Concat, [1]] # cat head P4- [-1, 2, C3k2, [512, False]] # 19 (P4/16-medium)- [-1, 1, Conv, [512, 3, 2]]- [[-1, 11], 1, Concat, [1]] # cat head P5- [-1, 2, C3k2, [1024, True]] # 22 (P5/32-large)- [[17, 20, 23], 1, Detect, [nc]] # Detect(P3, P4, P5)

以上添加位置仅供参考，具体添加位置以及模块效果以自己的数据集结果为准

3.2运行成功截图

OK 以上就是添加HAT注意力机制的全部过程了，后续将持续更新尽情期待