主页：HABUO🍁主页：HABUO
🍁YOLOv8入门+改进专栏🍁
🍁如果再也不能见到你，祝你早安，午安，晚安🍁
【YOLOv8改进系列】：
【YOLOv8】YOLOv8结构解读
YOLOv8改进系列（1）----替换主干网络之EfficientViT
YOLOv8改进系列（2）----替换主干网络之FasterNet
YOLOv8改进系列（3）----替换主干网络之ConvNeXt V2
YOLOv8改进系列（4）----替换C2f之FasterNet中的FasterBlock替换C2f中的Bottleneck
YOLOv8改进系列（5）----替换主干网络之EfficientFormerV2
YOLOv8改进系列（6）----替换主干网络之VanillaNet
YOLOv8改进系列（7）----替换主干网络之LSKNet
目录
💯一、Swin Transformer介绍
1. 简介
2. LSKNet架构设计
背景知识
研究方法
3. 实验与结果
4. 关键结论
💯二、具体添加方法
第①步：创建SwinTransformer.py
第②步：修改task.py
(1)引入创建的SwinTransformer文件
(2)修改_predict_once函数
(3)修改parse_model函数
第③步：yolov8.yaml文件修改
第④步：验证是否加入成功
💯一、Swin Transformer介绍
- 论文题目：《Swin Transformer: Hierarchical Vision Transformer using Shifted Windows》
- 论文地址：https://arxiv.org/pdf/2103.14030
1. 简介
论文介绍了一种新的视觉Transformer模型——Swin Transformer，它旨在成为计算机视觉领域的通用骨干网络。Swin Transformer通过其独特的层次化结构和移位窗口（Shifted Windows）机制，解决了传统Transformer在视觉任务中的计算复杂度问题，并在图像分类、目标检测和语义分割等多个任务中取得了优异的性能。
2. LSKNet架构设计
背景知识
传统的卷积神经网络（CNN）在计算机视觉领域占据主导地位，而Transformer架构在自然语言处理（NLP）中取得了巨大成功。然而，将Transformer直接应用于计算机视觉面临两大挑战：
-
视觉元素的尺度变化大：与语言中的固定尺度词元不同，视觉元素的尺度变化范围很大。
-
图像分辨率高：图像中的像素分辨率远高于文本中的单词，这使得Transformer在高分辨率图像上的计算复杂度呈二次方增长，难以处理密集预测任务。
研究方法
为了解决上述问题，Swin Transformer提出了以下创新点：
-
层次化Transformer结构：通过逐步合并小尺寸图像块（patches），构建层次化的特征图，从而能够处理不同尺度的视觉元素。
-
移位窗口机制：在计算自注意力时，将图像划分为非重叠的局部窗口，并在连续的Transformer块之间交替使用常规窗口划分和移位窗口划分。这种机制不仅保持了计算效率，还允许跨窗口连接，增强了模型的表达能力。
-
线性计算复杂度：通过在局部窗口内计算自注意力，Swin Transformer的计算复杂度与图像大小呈线性关系，使其适用于高分辨率图像和密集预测任务。
3. 实验与结果
Swin Transformer在多个计算机视觉任务上进行了广泛的实验，结果表明其性能显著优于现有的CNN和Transformer模型：
-
图像分类：在ImageNet-1K数据集上，Swin Transformer取得了87.3%的top-1准确率，超越了之前的最佳模型。
-
目标检测：在COCO数据集上，Swin Transformer取得了58.7 box AP和51.1 mask AP的性能，分别比之前的最佳结果高出+2.7 box AP和+2.6 mask AP。
-
语义分割：在ADE20K数据集上，Swin Transformer取得了53.5 mIoU的性能，比之前的最佳模型高出+3.2 mIoU。
4. 关键结论
Swin Transformer通过其层次化结构和移位窗口机制，有效地解决了传统Transformer在视觉任务中的计算复杂度问题，并在多个任务中取得了突破性的性能。其设计不仅适用于图像分类，还能很好地支持密集预测任务，如目标检测和语义分割。此外，Swin Transformer的线性计算复杂度使其能够处理高分辨率图像，为计算机视觉领域提供了一个强大的通用骨干网络。
💯二、具体添加方法
第①步：创建SwinTransformer.py
创建完成后，将下面代码直接复制粘贴进去：
import torch
import torch.nn as nn
import torch.nn.functional as F
import torch.utils.checkpoint as checkpoint
import numpy as np
from timm.models.layers import DropPath, to_2tuple, trunc_normal_
__all__ = ['SwinTransformer_Tiny']
class Mlp(nn.Module):
""" Multilayer perceptron."""
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)
"""
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):
""" 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
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):
""" Forward function.
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
class SwinTransformerBlock(nn.Module):
""" Swin Transformer Block.
Args:
dim (int): Number of input channels.
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, 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.num_heads = num_heads
self.window_size = window_size
self.shift_size = shift_size
self.mlp_ratio = mlp_ratio
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)
self.H = None
self.W = None
def forward(self, x, mask_matrix):
""" Forward function.
Args:
x: Input feature, tensor size (B, H*W, C).
H, W: Spatial resolution of the input feature.
mask_matrix: Attention mask for cyclic shift.
"""
B, L, C = x.shape
H, W = self.H, self.W
assert L == H * W, "input feature has wrong size"
shortcut = x
x = self.norm1(x)
x = x.view(B, H, W, C)
# pad feature maps to multiples of window size
pad_l = pad_t = 0
pad_r = (self.window_size - W % self.window_size) % self.window_size
pad_b = (self.window_size - H % self.window_size) % self.window_size
x = F.pad(x, (0, 0, pad_l, pad_r, pad_t, pad_b))
_, Hp, Wp, _ = x.shape
# cyclic shift
if self.shift_size > 0:
shifted_x = torch.roll(x, shifts=(-self.shift_size, -self.shift_size), dims=(1, 2))
attn_mask = mask_matrix.type(x.dtype)
else:
shifted_x = x
attn_mask = None
# 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=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, Hp, Wp) # 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
if pad_r > 0 or pad_b > 0:
x = x[:, :H, :W, :].contiguous()
x = x.view(B, H * W, C)
# FFN
x = shortcut + self.drop_path(x)
x = x + self.drop_path(self.mlp(self.norm2(x)))
return x
class PatchMerging(nn.Module):
""" Patch Merging Layer
Args:
dim (int): Number of input channels.
norm_layer (nn.Module, optional): Normalization layer. Default: nn.LayerNorm
"""
def __init__(self, dim, norm_layer=nn.LayerNorm):
super().__init__()
self.dim = dim
self.reduction = nn.Linear(4 * dim, 2 * dim, bias=False)
self.norm = norm_layer(4 * dim)
def forward(self, x, H, W):
""" Forward function.
Args:
x: Input feature, tensor size (B, H*W, C).
H, W: Spatial resolution of the input feature.
"""
B, L, C = x.shape
assert L == H * W, "input feature has wrong size"
x = x.view(B, H, W, C)
# padding
pad_input = (H % 2 == 1) or (W % 2 == 1)
if pad_input:
x = F.pad(x, (0, 0, 0, W % 2, 0, H % 2))
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
class BasicLayer(nn.Module):
""" A basic Swin Transformer layer for one stage.
Args:
dim (int): Number of feature channels
depth (int): Depths of this stage.
num_heads (int): Number of attention head.
window_size (int): Local window size. Default: 7.
mlp_ratio (float): Ratio of mlp hidden dim to embedding dim. Default: 4.
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,
depth,
num_heads,
window_size=7,
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.window_size = window_size
self.shift_size = window_size // 2
self.depth = depth
self.use_checkpoint = use_checkpoint
# build blocks
self.blocks = nn.ModuleList([
SwinTransformerBlock(
dim=dim,
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(dim=dim, norm_layer=norm_layer)
else:
self.downsample = None
def forward(self, x, H, W):
""" Forward function.
Args:
x: Input feature, tensor size (B, H*W, C).
H, W: Spatial resolution of the input feature.
"""
# calculate attention mask for SW-MSA
Hp = int(np.ceil(H / self.window_size)) * self.window_size
Wp = int(np.ceil(W / self.window_size)) * self.window_size
img_mask = torch.zeros((1, Hp, Wp, 1), device=x.device) # 1 Hp Wp 1
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))
for blk in self.blocks:
blk.H, blk.W = H, W
if self.use_checkpoint:
x = checkpoint.checkpoint(blk, x, attn_mask)
else:
x = blk(x, attn_mask)
if self.downsample is not None:
x_down = self.downsample(x, H, W)
Wh, Ww = (H + 1) // 2, (W + 1) // 2
return x, H, W, x_down, Wh, Ww
else:
return x, H, W, x, H, W
class PatchEmbed(nn.Module):
""" Image to Patch Embedding
Args:
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, patch_size=4, in_chans=3, embed_dim=96, norm_layer=None):
super().__init__()
patch_size = to_2tuple(patch_size)
self.patch_size = patch_size
self.in_chans = in_chans
self.embed_dim = embed_dim
self.proj = nn.Conv2d(in_chans, embed_dim, kernel_size=patch_size, stride=patch_size)
if norm_layer is not None:
self.norm = norm_layer(embed_dim)
else:
self.norm = None
def forward(self, x):
"""Forward function."""
# padding
_, _, H, W = x.size()
if W % self.patch_size[1] != 0:
x = F.pad(x, (0, self.patch_size[1] - W % self.patch_size[1]))
if H % self.patch_size[0] != 0:
x = F.pad(x, (0, 0, 0, self.patch_size[0] - H % self.patch_size[0]))
x = self.proj(x) # B C Wh Ww
if self.norm is not None:
Wh, Ww = x.size(2), x.size(3)
x = x.flatten(2).transpose(1, 2)
x = self.norm(x)
x = x.transpose(1, 2).view(-1, self.embed_dim, Wh, Ww)
return x
class SwinTransformer(nn.Module):
""" Swin Transformer backbone.
A PyTorch impl of : `Swin Transformer: Hierarchical Vision Transformer using Shifted Windows` -
https://arxiv.org/pdf/2103.14030
Args:
pretrain_img_size (int): Input image size for training the pretrained model,
used in absolute postion embedding. Default 224.
patch_size (int | tuple(int)): Patch size. Default: 4.
in_chans (int): Number of input image channels. Default: 3.
embed_dim (int): Number of linear projection output channels. Default: 96.
depths (tuple[int]): Depths of each Swin Transformer stage.
num_heads (tuple[int]): Number of attention head of each stage.
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.
drop_rate (float): Dropout rate.
attn_drop_rate (float): Attention dropout rate. Default: 0.
drop_path_rate (float): Stochastic depth rate. Default: 0.2.
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.
out_indices (Sequence[int]): Output from which stages.
frozen_stages (int): Stages to be frozen (stop grad and set eval mode).
-1 means not freezing any parameters.
use_checkpoint (bool): Whether to use checkpointing to save memory. Default: False.
"""
def __init__(self,
pretrain_img_size=224,
patch_size=4,
in_chans=3,
embed_dim=96,
depths=[2, 2, 6, 2],
num_heads=[3, 6, 12, 24],
window_size=7,
mlp_ratio=4.,
qkv_bias=True,
qk_scale=None,
drop_rate=0.,
attn_drop_rate=0.,
drop_path_rate=0.2,
norm_layer=nn.LayerNorm,
ape=False,
patch_norm=True,
out_indices=(0, 1, 2, 3),
frozen_stages=-1,
use_checkpoint=False):
super().__init__()
self.pretrain_img_size = pretrain_img_size
self.num_layers = len(depths)
self.embed_dim = embed_dim
self.ape = ape
self.patch_norm = patch_norm
self.out_indices = out_indices
self.frozen_stages = frozen_stages
# split image into non-overlapping patches
self.patch_embed = PatchEmbed(
patch_size=patch_size, in_chans=in_chans, embed_dim=embed_dim,
norm_layer=norm_layer if self.patch_norm else None)
# absolute position embedding
if self.ape:
pretrain_img_size = to_2tuple(pretrain_img_size)
patch_size = to_2tuple(patch_size)
patches_resolution = [pretrain_img_size[0] // patch_size[0], pretrain_img_size[1] // patch_size[1]]
self.absolute_pos_embed = nn.Parameter(torch.zeros(1, embed_dim, patches_resolution[0], patches_resolution[1]))
trunc_normal_(self.absolute_pos_embed, std=.02)
self.pos_drop = nn.Dropout(p=drop_rate)
# stochastic depth
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),
depth=depths[i_layer],
num_heads=num_heads[i_layer],
window_size=window_size,
mlp_ratio=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.append(layer)
num_features = [int(embed_dim * 2 ** i) for i in range(self.num_layers)]
self.num_features = num_features
# add a norm layer for each output
for i_layer in out_indices:
layer = norm_layer(num_features[i_layer])
layer_name = f'norm{i_layer}'
self.add_module(layer_name, layer)
self.channel = [i.size(1) for i in self.forward(torch.randn(1, 3, 640, 640))]
def forward(self, x):
"""Forward function."""
x = self.patch_embed(x)
Wh, Ww = x.size(2), x.size(3)
if self.ape:
# interpolate the position embedding to the corresponding size
absolute_pos_embed = F.interpolate(self.absolute_pos_embed, size=(Wh, Ww), mode='bicubic')
x = (x + absolute_pos_embed).flatten(2).transpose(1, 2) # B Wh*Ww C
else:
x = x.flatten(2).transpose(1, 2)
x = self.pos_drop(x)
outs = []
for i in range(self.num_layers):
layer = self.layers[i]
x_out, H, W, x, Wh, Ww = layer(x, Wh, Ww)
if i in self.out_indices:
norm_layer = getattr(self, f'norm{i}')
x_out = norm_layer(x_out)
out = x_out.view(-1, H, W, self.num_features[i]).permute(0, 3, 1, 2).contiguous()
outs.append(out)
return outs
def update_weight(model_dict, weight_dict):
idx, temp_dict = 0, {}
for k, v in weight_dict.items():
if k in model_dict.keys() and np.shape(model_dict[k]) == np.shape(v):
temp_dict[k] = v
idx += 1
model_dict.update(temp_dict)
print(f'loading weights... {idx}/{len(model_dict)} items')
return model_dict
def SwinTransformer_Tiny(weights=''):
model = SwinTransformer(depths=[2, 2, 6, 2], num_heads=[3, 6, 12, 24])
if weights:
model.load_state_dict(update_weight(model.state_dict(), torch.load(weights)['model']))
return model第②步：修改task.py
(1)引入创建的SwinTransformer文件
from ultralytics.nn.backbone.SwinTransformer import *(2)修改_predict_once函数
def _predict_once(self, x, profile=False, visualize=False, embed=None):
"""
Perform a forward pass through the network.
Args:
x (torch.Tensor): The input tensor to the model.
profile (bool): Print the computation time of each layer if True, defaults to False.
visualize (bool): Save the feature maps of the model if True, defaults to False.
embed (list, optional): A list of feature vectors/embeddings to return.
Returns:
(torch.Tensor): The last output of the model.
"""
y, dt, embeddings = [], [], [] # outputs
for idx, m in enumerate(self.model):
if m.f != -1: # if not from previous layer
x = y[m.f] if isinstance(m.f, int) else [x if j == -1 else y[j] for j in m.f] # from earlier layers
if profile:
self._profile_one_layer(m, x, dt)
if hasattr(m, 'backbone'):
x = m(x)
for _ in range(5 - len(x)):
x.insert(0, None)
for i_idx, i in enumerate(x):
if i_idx in self.save:
y.append(i)
else:
y.append(None)
# print(f'layer id:{idx:>2} {m.type:>50} output shape:{", ".join([str(x_.size()) for x_ in x if x_ is not None])}')
x = x[-1]
else:
x = m(x) # run
y.append(x if m.i in self.save else None) # save output
# if type(x) in {list, tuple}:
# if idx == (len(self.model) - 1):
# if type(x[1]) is dict:
# print(f'layer id:{idx:>2} {m.type:>50} output shape:{", ".join([str(x_.size()) for x_ in x[1]["one2one"]])}')
# else:
# print(f'layer id:{idx:>2} {m.type:>50} output shape:{", ".join([str(x_.size()) for x_ in x[1]])}')
# else:
# print(f'layer id:{idx:>2} {m.type:>50} output shape:{", ".join([str(x_.size()) for x_ in x if x_ is not None])}')
# elif type(x) is dict:
# print(f'layer id:{idx:>2} {m.type:>50} output shape:{", ".join([str(x_.size()) for x_ in x["one2one"]])}')
# else:
# if not hasattr(m, 'backbone'):
# print(f'layer id:{idx:>2} {m.type:>50} output shape:{x.size()}')
if visualize:
feature_visualization(x, m.type, m.i, save_dir=visualize)
if embed and m.i in embed:
embeddings.append(nn.functional.adaptive_avg_pool2d(x, (1, 1)).squeeze(-1).squeeze(-1)) # flatten
if m.i == max(embed):
return torch.unbind(torch.cat(embeddings, 1), dim=0)
return x(3)修改parse_model函数
可以直接把下面的代码粘贴到对应的位置中，后续的改进中，对应的模块就不需要做出改变，有改变处，后续会另有说明
def parse_model(d, ch, verbose=True, warehouse_manager=None): # model_dict, input_channels(3)
"""Parse a YOLO model.yaml dictionary into a PyTorch model."""
import ast
# Args
max_channels = float("inf")
nc, act, scales = (d.get(x) for x in ("nc", "activation", "scales"))
depth, width, kpt_shape = (d.get(x, 1.0) for x in ("depth_multiple", "width_multiple", "kpt_shape"))
if scales:
scale = d.get("scale")
if not scale:
scale = tuple(scales.keys())[0]
LOGGER.warning(f"WARNING ⚠️ no model scale passed. Assuming scale='{scale}'.")
if len(scales[scale]) == 3:
depth, width, max_channels = scales[scale]
elif len(scales[scale]) == 4:
depth, width, max_channels, threshold = scales[scale]
if act:
Conv.default_act = eval(act) # redefine default activation, i.e. Conv.default_act = nn.SiLU()
if verbose:
LOGGER.info(f"{colorstr('activation:')} {act}") # print
if verbose:
LOGGER.info(f"\n{'':>3}{'from':>20}{'n':>3}{'params':>10} {'module':<60}{'arguments':<50}")
ch = [ch]
layers, save, c2 = [], [], ch[-1] # layers, savelist, ch out
is_backbone = False
for i, (f, n, m, args) in enumerate(d["backbone"] + d["head"]): # from, number, module, args
try:
if m == 'node_mode':
m = d[m]
if len(args) > 0:
if args[0] == 'head_channel':
args[0] = int(d[args[0]])
t = m
m = getattr(torch.nn, m[3:]) if 'nn.' in m else globals()[m] # get module
except:
pass
for j, a in enumerate(args):
if isinstance(a, str):
with contextlib.suppress(ValueError):
try:
args[j] = locals()[a] if a in locals() else ast.literal_eval(a)
except:
args[j] = a
n = n_ = max(round(n * depth), 1) if n > 1 else n # depth gain
if m in {
Classify, Conv, ConvTranspose, GhostConv, Bottleneck, GhostBottleneck, SPP, SPPF, DWConv, Focus, BottleneckCSP, C1, C2, C2f, ELAN1, AConv, SPPELAN, C2fAttn, C3, C3TR,
C3Ghost, nn.Conv2d, nn.ConvTranspose2d, DWConvTranspose2d, C3x, RepC3, PSA, SCDown, C2fCIB, C2f_Faster, C2f_ODConv,
C2f_Faster_EMA, C2f_DBB, GSConv, GSConvns, VoVGSCSP, VoVGSCSPns, VoVGSCSPC, C2f_CloAtt, C3_CloAtt, SCConv, C2f_SCConv, C3_SCConv, C2f_ScConv, C3_ScConv,
C3_EMSC, C3_EMSCP, C2f_EMSC, C2f_EMSCP, RCSOSA, KWConv, C2f_KW, C3_KW, DySnakeConv, C2f_DySnakeConv, C3_DySnakeConv,
DCNv2, C3_DCNv2, C2f_DCNv2, DCNV3_YOLO, C3_DCNv3, C2f_DCNv3, C3_Faster, C3_Faster_EMA, C3_ODConv,
OREPA, OREPA_LargeConv, RepVGGBlock_OREPA, C3_OREPA, C2f_OREPA, C3_DBB, C3_REPVGGOREPA, C2f_REPVGGOREPA,
C3_DCNv2_Dynamic, C2f_DCNv2_Dynamic, C3_ContextGuided, C2f_ContextGuided, C3_MSBlock, C2f_MSBlock,
C3_DLKA, C2f_DLKA, CSPStage, SPDConv, RepBlock, C3_EMBC, C2f_EMBC, SPPF_LSKA, C3_DAttention, C2f_DAttention,
C3_Parc, C2f_Parc, C3_DWR, C2f_DWR, RFAConv, RFCAConv, RFCBAMConv, C3_RFAConv, C2f_RFAConv,
C3_RFCBAMConv, C2f_RFCBAMConv, C3_RFCAConv, C2f_RFCAConv, C3_FocusedLinearAttention, C2f_FocusedLinearAttention,
C3_AKConv, C2f_AKConv, AKConv, C3_MLCA, C2f_MLCA,
C3_UniRepLKNetBlock, C2f_UniRepLKNetBlock, C3_DRB, C2f_DRB, C3_DWR_DRB, C2f_DWR_DRB, CSP_EDLAN,
C3_AggregatedAtt, C2f_AggregatedAtt, DCNV4_YOLO, C3_DCNv4, C2f_DCNv4, HWD, SEAM,
C3_SWC, C2f_SWC, C3_iRMB, C2f_iRMB, C3_iRMB_Cascaded, C2f_iRMB_Cascaded, C3_iRMB_DRB, C2f_iRMB_DRB, C3_iRMB_SWC, C2f_iRMB_SWC,
C3_VSS, C2f_VSS, C3_LVMB, C2f_LVMB, RepNCSPELAN4, DBBNCSPELAN4, OREPANCSPELAN4, DRBNCSPELAN4, ADown, V7DownSampling,
C3_DynamicConv, C2f_DynamicConv, C3_GhostDynamicConv, C2f_GhostDynamicConv, C3_RVB, C2f_RVB, C3_RVB_SE, C2f_RVB_SE, C3_RVB_EMA, C2f_RVB_EMA, DGCST,
C3_RetBlock, C2f_RetBlock, C3_PKIModule, C2f_PKIModule, RepNCSPELAN4_CAA, C3_FADC, C2f_FADC, C3_PPA, C2f_PPA, SRFD, DRFD, RGCSPELAN,
C3_Faster_CGLU, C2f_Faster_CGLU, C3_Star, C2f_Star, C3_Star_CAA, C2f_Star_CAA, C3_KAN, C2f_KAN, C3_EIEM, C2f_EIEM, C3_DEConv, C2f_DEConv,
C3_SMPCGLU, C2f_SMPCGLU, C3_Heat, C2f_Heat, CSP_PTB, SimpleStem, VisionClueMerge, VSSBlock_YOLO, XSSBlock, GLSA, C2f_WTConv, WTConv2d, FeaturePyramidSharedConv,
C2f_FMB, LDConv, C2f_gConv, C2f_WDBB, C2f_DeepDBB, C2f_AdditiveBlock, C2f_AdditiveBlock_CGLU, CSP_MSCB, C2f_MSMHSA_CGLU, CSP_PMSFA, C2f_MogaBlock,
C2f_SHSA, C2f_SHSA_CGLU, C2f_SMAFB, C2f_SMAFB_CGLU, C2f_IdentityFormer, C2f_RandomMixing, C2f_PoolingFormer, C2f_ConvFormer, C2f_CaFormer,
C2f_IdentityFormerCGLU, C2f_RandomMixingCGLU, C2f_PoolingFormerCGLU, C2f_ConvFormerCGLU, C2f_CaFormerCGLU, CSP_MutilScaleEdgeInformationEnhance, C2f_FFCM,
C2f_SFHF, CSP_FreqSpatial, C2f_MSM, C2f_RAB, C2f_HDRAB, C2f_LFE, CSP_MutilScaleEdgeInformationSelect, C2f_SFA, C2f_CTA, C2f_CAMixer, MANet,
MANet_FasterBlock, MANet_FasterCGLU, MANet_Star, C2f_HFERB, C2f_DTAB, C2f_ETB, C2f_JDPM, C2f_AP, PSConv, C2f_Kat, C2f_Faster_KAN, C2f_Strip, C2f_StripCGLU
}:
if args[0] == 'head_channel':
args[0] = d[args[0]]
c1, c2 = ch[f], args[0]
if c2 != nc: # if c2 not equal to number of classes (i.e. for Classify() output)
c2 = make_divisible(min(c2, max_channels) * width, 8)
if m is C2fAttn:
args[1] = make_divisible(min(args[1], max_channels // 2) * width, 8) # embed channels
args[2] = int(
max(round(min(args[2], max_channels // 2 // 32)) * width, 1) if args[2] > 1 else args[2]
) # num heads
args = [c1, c2, *args[1:]]
if m in (KWConv, C2f_KW, C3_KW):
args.insert(2, f'layer{i}')
args.insert(2, warehouse_manager)
if m in (DySnakeConv,):
c2 = c2 * 3
if m in (RepNCSPELAN4, DBBNCSPELAN4, OREPANCSPELAN4, DRBNCSPELAN4, RepNCSPELAN4_CAA):
args[2] = make_divisible(min(args[2], max_channels) * width, 8)
args[3] = make_divisible(min(args[3], max_channels) * width, 8)
if m in {
BottleneckCSP, C1, C2, C2f, C2fAttn, C3, C3TR, C3Ghost, C3x, RepC3, C2fCIB, C2f_Faster, C2f_ODConv, C2f_Faster_EMA, C2f_DBB,
VoVGSCSP, VoVGSCSPns, VoVGSCSPC, C2f_CloAtt, C3_CloAtt, C2f_SCConv, C3_SCConv, C2f_ScConv, C3_ScConv,
C3_EMSC, C3_EMSCP, C2f_EMSC, C2f_EMSCP, RCSOSA, C2f_KW, C3_KW, C2f_DySnakeConv, C3_DySnakeConv,
C3_DCNv2, C2f_DCNv2, C3_DCNv3, C2f_DCNv3, C3_Faster, C3_Faster_EMA, C3_ODConv, C3_OREPA, C2f_OREPA, C3_DBB,
C3_REPVGGOREPA, C2f_REPVGGOREPA, C3_DCNv2_Dynamic, C2f_DCNv2_Dynamic, C3_ContextGuided, C2f_ContextGuided,
C3_MSBlock, C2f_MSBlock, C3_DLKA, C2f_DLKA, CSPStage, RepBlock, C3_EMBC, C2f_EMBC, C3_DAttention, C2f_DAttention,
C3_Parc, C2f_Parc, C3_DWR, C2f_DWR, C3_RFAConv, C2f_RFAConv, C3_RFCBAMConv, C2f_RFCBAMConv, C3_RFCAConv, C2f_RFCAConv,
C3_FocusedLinearAttention, C2f_FocusedLinearAttention, C3_AKConv, C2f_AKConv, C3_MLCA, C2f_MLCA,
C3_UniRepLKNetBlock, C2f_UniRepLKNetBlock, C3_DRB, C2f_DRB, C3_DWR_DRB, C2f_DWR_DRB, CSP_EDLAN,
C3_AggregatedAtt, C2f_AggregatedAtt, C3_DCNv4, C2f_DCNv4, C3_SWC, C2f_SWC,
C3_iRMB, C2f_iRMB, C3_iRMB_Cascaded, C2f_iRMB_Cascaded, C3_iRMB_DRB, C2f_iRMB_DRB, C3_iRMB_SWC, C2f_iRMB_SWC,
C3_VSS, C2f_VSS, C3_LVMB, C2f_LVMB, C3_DynamicConv, C2f_DynamicConv, C3_GhostDynamicConv, C2f_GhostDynamicConv,
C3_RVB, C2f_RVB, C3_RVB_SE, C2f_RVB_SE, C3_RVB_EMA, C2f_RVB_EMA, C3_RetBlock, C2f_RetBlock, C3_PKIModule, C2f_PKIModule,
C3_FADC, C2f_FADC, C3_PPA, C2f_PPA, RGCSPELAN, C3_Faster_CGLU, C2f_Faster_CGLU, C3_Star, C2f_Star, C3_Star_CAA, C2f_Star_CAA,
C3_KAN, C2f_KAN, C3_EIEM, C2f_EIEM, C3_DEConv, C2f_DEConv, C3_SMPCGLU, C2f_SMPCGLU, C3_Heat, C2f_Heat, CSP_PTB, XSSBlock, C2f_WTConv,
C2f_FMB, C2f_gConv, C2f_WDBB, C2f_DeepDBB, C2f_AdditiveBlock, C2f_AdditiveBlock_CGLU, CSP_MSCB, C2f_MSMHSA_CGLU, CSP_PMSFA, C2f_MogaBlock,
C2f_SHSA, C2f_SHSA_CGLU, C2f_SMAFB, C2f_SMAFB_CGLU, C2f_IdentityFormer, C2f_RandomMixing, C2f_PoolingFormer, C2f_ConvFormer, C2f_CaFormer,
C2f_IdentityFormerCGLU, C2f_RandomMixingCGLU, C2f_PoolingFormerCGLU, C2f_ConvFormerCGLU, C2f_CaFormerCGLU, CSP_MutilScaleEdgeInformationEnhance, C2f_FFCM,
C2f_SFHF, CSP_FreqSpatial, C2f_MSM, C2f_RAB, C2f_HDRAB, C2f_LFE, CSP_MutilScaleEdgeInformationSelect, C2f_SFA, C2f_CTA, C2f_CAMixer, MANet,
MANet_FasterBlock, MANet_FasterCGLU, MANet_Star, C2f_HFERB, C2f_DTAB, C2f_ETB, C2f_JDPM, C2f_AP, C2f_Kat, C2f_Faster_KAN, C2f_Strip, C2f_StripCGLU
}:
args.insert(2, n) # number of repeats
n = 1
elif m in {AIFI, AIFI_RepBN}:
args = [ch[f], *args]
c2 = args[0]
elif m in (HGStem, HGBlock, Ghost_HGBlock, Rep_HGBlock, Dynamic_HGBlock, EIEStem):
c1, cm, c2 = ch[f], args[0], args[1]
if c2 != nc: # if c2 not equal to number of classes (i.e. for Classify() output)
c2 = make_divisible(min(c2, max_channels) * width, 8)
cm = make_divisible(min(cm, max_channels) * width, 8)
args = [c1, cm, c2, *args[2:]]
if m in (HGBlock, Ghost_HGBlock, Rep_HGBlock, Dynamic_HGBlock):
args.insert(4, n) # number of repeats
n = 1
elif m is ResNetLayer:
c2 = args[1] if args[3] else args[1] * 4
elif m is nn.BatchNorm2d:
args = [ch[f]]
elif m is Concat:
c2 = sum(ch[x] for x in f)
elif m in ((WorldDetect, ImagePoolingAttn) + DETECT_CLASS + V10_DETECT_CLASS + SEGMENT_CLASS + POSE_CLASS + OBB_CLASS):
args.append([ch[x] for x in f])
if m in SEGMENT_CLASS:
args[2] = make_divisible(min(args[2], max_channels) * width, 8)
if m in (Segment_LSCD, Segment_TADDH, Segment_LSCSBD, Segment_LSDECD, Segment_RSCD):
args[3] = make_divisible(min(args[3], max_channels) * width, 8)
if m in (Detect_LSCD, Detect_TADDH, Detect_LSCSBD, Detect_LSDECD, Detect_RSCD, v10Detect_LSCD, v10Detect_TADDH, v10Detect_RSCD, v10Detect_LSDECD):
args[1] = make_divisible(min(args[1], max_channels) * width, 8)
if m in (Pose_LSCD, Pose_TADDH, Pose_LSCSBD, Pose_LSDECD, Pose_RSCD, OBB_LSCD, OBB_TADDH, OBB_LSCSBD, OBB_LSDECD, OBB_RSCD):
args[2] = make_divisible(min(args[2], max_channels) * width, 8)
elif m is RTDETRDecoder: # special case, channels arg must be passed in index 1
args.insert(1, [ch[x] for x in f])
elif m is Fusion:
args[0] = d[args[0]]
c1, c2 = [ch[x] for x in f], (sum([ch[x] for x in f]) if args[0] == 'concat' else ch[f[0]])
args = [c1, args[0]]
elif m is CBLinear:
c2 = make_divisible(min(args[0][-1], max_channels) * width, 8)
c1 = ch[f]
args = [c1, [make_divisible(min(c2_, max_channels) * width, 8) for c2_ in args[0]], *args[1:]]
elif m is CBFuse:
c2 = ch[f[-1]]
elif isinstance(m, str):
t = m
if len(args) == 2:
m = timm.create_model(m, pretrained=args[0], pretrained_cfg_overlay={'file':args[1]}, features_only=True)
elif len(args) == 1:
m = timm.create_model(m, pretrained=args[0], features_only=True)
c2 = m.feature_info.channels()
elif m in {convnextv2_atto, convnextv2_femto, convnextv2_pico, convnextv2_nano, convnextv2_tiny, convnextv2_base, convnextv2_large, convnextv2_huge,
fasternet_t0, fasternet_t1, fasternet_t2, fasternet_s, fasternet_m, fasternet_l,
EfficientViT_M0, EfficientViT_M1, EfficientViT_M2, EfficientViT_M3, EfficientViT_M4, EfficientViT_M5,
efficientformerv2_s0, efficientformerv2_s1, efficientformerv2_s2, efficientformerv2_l,
vanillanet_5, vanillanet_6, vanillanet_7, vanillanet_8, vanillanet_9, vanillanet_10, vanillanet_11, vanillanet_12, vanillanet_13, vanillanet_13_x1_5, vanillanet_13_x1_5_ada_pool,
RevCol,
lsknet_t, lsknet_s,
SwinTransformer_Tiny,
repvit_m0_9, repvit_m1_0, repvit_m1_1, repvit_m1_5, repvit_m2_3,
CSWin_tiny, CSWin_small, CSWin_base, CSWin_large,
unireplknet_a, unireplknet_f, unireplknet_p, unireplknet_n, unireplknet_t, unireplknet_s, unireplknet_b, unireplknet_l, unireplknet_xl,
transnext_micro, transnext_tiny, transnext_small, transnext_base,
RMT_T, RMT_S, RMT_B, RMT_L,
PKINET_T, PKINET_S, PKINET_B,
MobileNetV4ConvSmall, MobileNetV4ConvMedium, MobileNetV4ConvLarge, MobileNetV4HybridMedium, MobileNetV4HybridLarge,
starnet_s050, starnet_s100, starnet_s150, starnet_s1, starnet_s2, starnet_s3, starnet_s4
}:
if m is RevCol:
args[1] = [make_divisible(min(k, max_channels) * width, 8) for k in args[1]]
args[2] = [max(round(k * depth), 1) for k in args[2]]
m = m(*args)
c2 = m.channel
elif m in {EMA, SpatialAttention, BiLevelRoutingAttention, BiLevelRoutingAttention_nchw,
TripletAttention, CoordAtt, CBAM, BAMBlock, LSKBlock, ScConv, LAWDS, EMSConv, EMSConvP,
SEAttention, CPCA, Partial_conv3, FocalModulation, EfficientAttention, MPCA, deformable_LKA,
EffectiveSEModule, LSKA, SegNext_Attention, DAttention, MLCA, TransNeXt_AggregatedAttention,
FocusedLinearAttention, LocalWindowAttention, ChannelAttention_HSFPN, ELA_HSFPN, CA_HSFPN, CAA_HSFPN,
DySample, CARAFE, CAA, ELA, CAFM, AFGCAttention, EUCB, ContrastDrivenFeatureAggregation, FSA}:
c2 = ch[f]
args = [c2, *args]
# print(args)
elif m in {SimAM, SpatialGroupEnhance}:
c2 = ch[f]
elif m is ContextGuidedBlock_Down:
c2 = ch[f] * 2
args = [ch[f], c2, *args]
elif m is BiFusion:
c1 = [ch[x] for x in f]
c2 = make_divisible(min(args[0], max_channels) * width, 8)
args = [c1, c2]
# --------------GOLD-YOLO--------------
elif m in {SimFusion_4in, AdvPoolFusion}:
c2 = sum(ch[x] for x in f)
elif m is SimFusion_3in:
c2 = args[0]
if c2 != nc: # if c2 not equal to number of classes (i.e. for Classify() output)
c2 = make_divisible(min(c2, max_channels) * width, 8)
args = [[ch[f_] for f_ in f], c2]
elif m is IFM:
c1 = ch[f]
c2 = sum(args[0])
args = [c1, *args]
elif m is InjectionMultiSum_Auto_pool:
c1 = ch[f[0]]
c2 = args[0]
args = [c1, *args]
elif m is PyramidPoolAgg:
c2 = args[0]
args = [sum([ch[f_] for f_ in f]), *args]
elif m is TopBasicLayer:
c2 = sum(args[1])
# --------------GOLD-YOLO--------------
# --------------ASF--------------
elif m is Zoom_cat:
c2 = sum(ch[x] for x in f)
elif m is Add:
c2 = ch[f[-1]]
elif m in {ScalSeq, DynamicScalSeq}:
c1 = [ch[x] for x in f]
c2 = make_divisible(args[0] * width, 8)
args = [c1, c2]
elif m is asf_attention_model:
args = [ch[f[-1]]]
# --------------ASF--------------
elif m is SDI:
args = [[ch[x] for x in f]]
elif m is Multiply:
c2 = ch[f[0]]
elif m is FocusFeature:
c1 = [ch[x] for x in f]
c2 = int(c1[1] * 0.5 * 3)
args = [c1, *args]
elif m is DASI:
c1 = [ch[x] for x in f]
args = [c1, c2]
elif m is CSMHSA:
c1 = [ch[x] for x in f]
c2 = ch[f[-1]]
args = [c1, c2]
elif m is CFC_CRB:
c1 = ch[f]
c2 = c1 // 2
args = [c1, *args]
elif m is SFC_G2:
c1 = [ch[x] for x in f]
c2 = c1[0]
args = [c1]
elif m in {CGAFusion, CAFMFusion, SDFM, PSFM}:
c2 = ch[f[1]]
args = [c2, *args]
elif m in {ContextGuideFusionModule}:
c1 = [ch[x] for x in f]
c2 = 2 * c1[1]
args = [c1]
# elif m in {PSA}:
# c2 = ch[f]
# args = [c2, *args]
elif m in {SBA}:
c1 = [ch[x] for x in f]
c2 = c1[-1]
args = [c1, c2]
elif m in {WaveletPool}:
c2 = ch[f] * 4
elif m in {WaveletUnPool}:
c2 = ch[f] // 4
elif m in {CSPOmniKernel}:
c2 = ch[f]
args = [c2]
elif m in {ChannelTransformer, PyramidContextExtraction}:
c1 = [ch[x] for x in f]
c2 = c1
args = [c1]
elif m in {RCM}:
c2 = ch[f]
args = [c2, *args]
elif m in {DynamicInterpolationFusion}:
c2 = ch[f[0]]
args = [[ch[x] for x in f]]
elif m in {FuseBlockMulti}:
c2 = ch[f[0]]
args = [c2]
elif m in {CrossLayerChannelAttention, CrossLayerSpatialAttention}:
c2 = [ch[x] for x in f]
args = [c2[0], *args]
elif m in {FreqFusion}:
c2 = ch[f[0]]
args = [[ch[x] for x in f], *args]
elif m in {DynamicAlignFusion}:
c2 = args[0]
args = [[ch[x] for x in f], c2]
elif m in {ConvEdgeFusion}:
c2 = make_divisible(min(args[0], max_channels) * width, 8)
args = [[ch[x] for x in f], c2]
elif m in {MutilScaleEdgeInfoGenetator}:
c1 = ch[f]
c2 = [make_divisible(min(i, max_channels) * width, 8) for i in args[0]]
args = [c1, c2]
elif m in {MultiScaleGatedAttn}:
c1 = [ch[x] for x in f]
c2 = min(c1)
args = [c1]
elif m in {WFU, MultiScalePCA, MultiScalePCA_Down}:
c1 = [ch[x] for x in f]
c2 = c1[0]
args = [c1]
elif m in {GetIndexOutput}:
c2 = ch[f][args[0]]
elif m is HyperComputeModule:
c1, c2 = ch[f], args[0]
c2 = make_divisible(min(c2, max_channels) * width, 8)
args = [c1, c2, threshold]
else:
c2 = ch[f]
if isinstance(c2, list) and m not in {ChannelTransformer, PyramidContextExtraction, CrossLayerChannelAttention, CrossLayerSpatialAttention, MutilScaleEdgeInfoGenetator}:
is_backbone = True
m_ = m
m_.backbone = True
else:
m_ = nn.Sequential(*(m(*args) for _ in range(n))) if n > 1 else m(*args) # module
t = str(m)[8:-2].replace('__main__.', '') # module type
m.np = sum(x.numel() for x in m_.parameters()) # number params
m_.i, m_.f, m_.type = i + 4 if is_backbone else i, f, t # attach index, 'from' index, type
if verbose:
LOGGER.info(f"{i:>3}{str(f):>20}{n_:>3}{m.np:10.0f} {t:<60}{str(args):<50}") # print
save.extend(x % (i + 4 if is_backbone else i) for x in ([f] if isinstance(f, int) else f) if x != -1) # append to savelist
layers.append(m_)
if i == 0:
ch = []
if isinstance(c2, list) and m not in {ChannelTransformer, PyramidContextExtraction, CrossLayerChannelAttention, CrossLayerSpatialAttention, MutilScaleEdgeInfoGenetator}:
ch.extend(c2)
for _ in range(5 - len(ch)):
ch.insert(0, 0)
else:
ch.append(c2)
return nn.Sequential(*layers), sorted(save)第③步：yolov8.yaml文件修改
在下述文件夹中创立yolov8-swintransformer.yaml
# Parameters
nc: 80 # number of classes
scales: # model compound scaling constants, i.e. 'model=yolov8n.yaml' will call yolov8.yaml with scale 'n'
# [depth, width, max_channels]
n: [0.33, 0.25, 1024] # YOLOv8n summary: 225 layers, 3157200 parameters, 3157184 gradients, 8.9 GFLOPs
s: [0.33, 0.50, 1024] # YOLOv8s summary: 225 layers, 11166560 parameters, 11166544 gradients, 28.8 GFLOPs
m: [0.67, 0.75, 768] # YOLOv8m summary: 295 layers, 25902640 parameters, 25902624 gradients, 79.3 GFLOPs
l: [1.00, 1.00, 512] # YOLOv8l summary: 365 layers, 43691520 parameters, 43691504 gradients, 165.7 GFLOPs
x: [1.00, 1.25, 512] # YOLOv8x summary: 365 layers, 68229648 parameters, 68229632 gradients, 258.5 GFLOPs
# 0-P1/2
# 1-P2/4
# 2-P3/8
# 3-P4/16
# 4-P5/32
# YOLOv8.0n backbone
backbone:
# [from, repeats, module, args]
- [-1, 1, SwinTransformer_Tiny, []] # 4
- [-1, 1, SPPF, [1024, 5]] # 5
# YOLOv8.0n head
head:
- [-1, 1, nn.Upsample, [None, 2, 'nearest']] # 6
- [[-1, 3], 1, Concat, [1]] # 7 cat backbone P4
- [-1, 3, C2f, [512]] # 8
- [-1, 1, nn.Upsample, [None, 2, 'nearest']] # 9
- [[-1, 2], 1, Concat, [1]] # 10 cat backbone P3
- [-1, 3, C2f, [256]] # 11 (P3/8-small)
- [-1, 1, Conv, [256, 3, 2]] # 12
- [[-1, 8], 1, Concat, [1]] # 13 cat head P4
- [-1, 3, C2f, [512]] # 14 (P4/16-medium)
- [-1, 1, Conv, [512, 3, 2]] # 15
- [[-1, 5], 1, Concat, [1]] # 16 cat head P5
- [-1, 3, C2f, [1024]] # 17 (P5/32-large)
- [[11, 14, 17], 1, Detect, [nc]] # Detect(P3, P4, P5)第④步：验证是否加入成功
将train.py中的配置文件进行修改，并运行
🏋不是每一粒种子都能开花，但播下种子就比荒芜的旷野强百倍🏋
🍁YOLOv8入门+改进专栏🍁
【YOLOv8改进系列】：
【YOLOv8】YOLOv8结构解读
YOLOv8改进系列（1）----替换主干网络之EfficientViT
YOLOv8改进系列（2）----替换主干网络之FasterNet
YOLOv8改进系列（3）----替换主干网络之ConvNeXt V2
YOLOv8改进系列（4）----替换C2f之FasterNet中的FasterBlock替换C2f中的Bottleneck
YOLOv8改进系列（5）----替换主干网络之EfficientFormerV2
YOLOv8改进系列（6）----替换主干网络之VanillaNet
YOLOv8改进系列（7）----替换主干网络之LSKNet
