目标检测算法改进系列之Backbone替换为RIFormer

RIFormer简介

Token Mixer是ViT骨干非常重要的组成成分，它用于对不同空域位置信息进行自适应聚合，但常规的自注意力往往存在高计算复杂度与高延迟问题。而直接移除Token Mixer又会导致不完备的结构先验，进而导致严重的性能下降。

原文地址：RIFormer: Keep Your Vision Backbone Effective But Removing Token Mixer

基于此，本文基于重参数机制提出了RepIdentityFormer方案以研究无Token Mixer的架构体系。紧接着，作者改进了学习架构以打破无Token Mixer架构的局限性并总结了5条指导方针。搭配上所提优化策略后，本文构建了一种极致简单且具有优异性能的视觉骨干，此外它还具有高推理效率优势。

实验结果表明：通过合适的优化策略，网络结构的归纳偏置可以被集成进简单架构体系中。本文为后续优化驱动的高效网络设计提供了新的起点和思路。

RIFormer代码实现

# Copyright (c) OpenMMLab. All rights reserved.
from typing import Sequence
import torch
import torch.nn as nn
import numpy as np
from mmcv.cnn.bricks import DropPath, build_activation_layer, build_norm_layer
from mmengine.model import BaseModule

__all__ = ['RIFormer']

class Mlp(nn.Module):
    """Mlp implemented by with 1*1 convolutions.

    Input: Tensor with shape [B, C, H, W].
    Output: Tensor with shape [B, C, H, W].
    Args:
        in_features (int): Dimension of input features.
        hidden_features (int): Dimension of hidden features.
        out_features (int): Dimension of output features.
        act_cfg (dict): The config dict for activation between pointwise
            convolution. Defaults to ``dict(type='GELU')``.
        drop (float): Dropout rate. Defaults to 0.0.
    """

    def __init__(self,
                 in_features,
                 hidden_features=None,
                 out_features=None,
                 act_cfg=dict(type='GELU'),
                 drop=0.):
        super().__init__()
        out_features = out_features or in_features
        hidden_features = hidden_features or in_features
        self.fc1 = nn.Conv2d(in_features, hidden_features, 1)
        self.act = build_activation_layer(act_cfg)
        self.fc2 = nn.Conv2d(hidden_features, out_features, 1)
        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

class PatchEmbed(nn.Module):
    """Patch Embedding module implemented by a layer of convolution.

    Input: tensor in shape [B, C, H, W]
    Output: tensor in shape [B, C, H/stride, W/stride]
    Args:
        patch_size (int): Patch size of the patch embedding. Defaults to 16.
        stride (int): Stride of the patch embedding. Defaults to 16.
        padding (int): Padding of the patch embedding. Defaults to 0.
        in_chans (int): Input channels. Defaults to 3.
        embed_dim (int): Output dimension of the patch embedding.
            Defaults to 768.
        norm_layer (module): Normalization module. Defaults to None (not use).
    """

    def __init__(self,
                 patch_size=16,
                 stride=16,
                 padding=0,
                 in_chans=3,
                 embed_dim=768,
                 norm_layer=None):
        super().__init__()
        self.proj = nn.Conv2d(
            in_chans,
            embed_dim,
            kernel_size=patch_size,
            stride=stride,
            padding=padding)
        self.norm = norm_layer(embed_dim) if norm_layer else nn.Identity()

    def forward(self, x):
        x = self.proj(x)
        x = self.norm(x)
        return x


class Affine(nn.Module):
    """Affine Transformation module.

    Args:
        in_features (int): Input dimension.
    """

    def __init__(self, in_features):
        super().__init__()
        self.affine = nn.Conv2d(
            in_features,
            in_features,
            kernel_size=1,
            stride=1,
            padding=0,
            groups=in_features,
            bias=True)

    def forward(self, x):
        return self.affine(x) - x


class RIFormerBlock(BaseModule):
    """RIFormer Block.

    Args:
        dim (int): Embedding dim.
        mlp_ratio (float): Mlp expansion ratio. Defaults to 4.
        norm_cfg (dict): The config dict for norm layers.
            Defaults to ``dict(type='GN', num_groups=1)``.
        act_cfg (dict): The config dict for activation between pointwise
            convolution. Defaults to ``dict(type='GELU')``.
        drop (float): Dropout rate. Defaults to 0.
        drop_path (float): Stochastic depth rate. Defaults to 0.
        layer_scale_init_value (float): Init value for Layer Scale.
            Defaults to 1e-5.
        deploy (bool): Whether to switch the model structure to
            deployment mode. Default: False.
    """

    def __init__(self,
                 dim,
                 mlp_ratio=4.,
                 norm_cfg=dict(type='GN', num_groups=1),
                 act_cfg=dict(type='GELU'),
                 drop=0.,
                 drop_path=0.,
                 layer_scale_init_value=1e-5,
                 deploy=False):

        super().__init__()

        if deploy:
            self.norm_reparam = build_norm_layer(norm_cfg, dim)[1]
        else:
            self.norm1 = build_norm_layer(norm_cfg, dim)[1]
            self.token_mixer = Affine(in_features=dim)
        self.norm2 = build_norm_layer(norm_cfg, dim)[1]
        mlp_hidden_dim = int(dim * mlp_ratio)
        self.mlp = Mlp(
            in_features=dim,
            hidden_features=mlp_hidden_dim,
            act_cfg=act_cfg,
            drop=drop)

        # The following two techniques are useful to train deep RIFormers.
        self.drop_path = DropPath(drop_path) if drop_path > 0. \
            else nn.Identity()
        self.layer_scale_1 = nn.Parameter(
            layer_scale_init_value * torch.ones((dim)), requires_grad=True)
        self.layer_scale_2 = nn.Parameter(
            layer_scale_init_value * torch.ones((dim)), requires_grad=True)
        self.norm_cfg = norm_cfg
        self.dim = dim
        self.deploy = deploy

    def forward(self, x):
        if hasattr(self, 'norm_reparam'):
            x = x + self.drop_path(
                self.layer_scale_1.unsqueeze(-1).unsqueeze(-1) *
                self.norm_reparam(x))
            x = x + self.drop_path(
                self.layer_scale_2.unsqueeze(-1).unsqueeze(-1) *
                self.mlp(self.norm2(x)))
        else:
            x = x + self.drop_path(
                self.layer_scale_1.unsqueeze(-1).unsqueeze(-1) *
                self.token_mixer(self.norm1(x)))
            x = x + self.drop_path(
                self.layer_scale_2.unsqueeze(-1).unsqueeze(-1) *
                self.mlp(self.norm2(x)))
        return x

    def fuse_affine(self, norm, token_mixer):
        gamma_affn = token_mixer.affine.weight.reshape(-1)
        gamma_affn = gamma_affn - torch.ones_like(gamma_affn)
        beta_affn = token_mixer.affine.bias
        gamma_ln = norm.weight
        beta_ln = norm.bias
        return (gamma_ln * gamma_affn), (beta_ln * gamma_affn + beta_affn)

    def get_equivalent_scale_bias(self):
        eq_s, eq_b = self.fuse_affine(self.norm1, self.token_mixer)
        return eq_s, eq_b

    def switch_to_deploy(self):
        if self.deploy:
            return
        eq_s, eq_b = self.get_equivalent_scale_bias()
        self.norm_reparam = build_norm_layer(self.norm_cfg, self.dim)[1]
        self.norm_reparam.weight.data = eq_s
        self.norm_reparam.bias.data = eq_b
        self.__delattr__('norm1')
        if hasattr(self, 'token_mixer'):
            self.__delattr__('token_mixer')
        self.deploy = True


def basic_blocks(dim,
                 index,
                 layers,
                 mlp_ratio=4.,
                 norm_cfg=dict(type='GN', num_groups=1),
                 act_cfg=dict(type='GELU'),
                 drop_rate=.0,
                 drop_path_rate=0.,
                 layer_scale_init_value=1e-5,
                 deploy=False):
    """generate RIFormer blocks for a stage."""
    blocks = []
    for block_idx in range(layers[index]):
        block_dpr = drop_path_rate * (block_idx + sum(layers[:index])) / (
            sum(layers) - 1)
        blocks.append(
            RIFormerBlock(
                dim,
                mlp_ratio=mlp_ratio,
                norm_cfg=norm_cfg,
                act_cfg=act_cfg,
                drop=drop_rate,
                drop_path=block_dpr,
                layer_scale_init_value=layer_scale_init_value,
                deploy=deploy,
            ))
    blocks = nn.Sequential(*blocks)

    return blocks

def update_weight(model_dict, weight_dict):
    idx, temp_dict = 0, {}
    for k, v in weight_dict.items():
        k = k[9:]
        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

class RIFormer(nn.Module):
    """RIFormer.

    A PyTorch implementation of RIFormer introduced by:
    `RIFormer: Keep Your Vision Backbone Effective But Removing Token Mixer <https://arxiv.org/abs/xxxx.xxxxx>`_

    Args:
        arch (str | dict): The model's architecture. If string, it should be
            one of architecture in ``RIFormer.arch_settings``. And if dict, it
            should include the following two keys:

            - layers (list[int]): Number of blocks at each stage.
            - embed_dims (list[int]): The number of channels at each stage.
            - mlp_ratios (list[int]): Expansion ratio of MLPs.
            - layer_scale_init_value (float): Init value for Layer Scale.

            Defaults to 'S12'.

        norm_cfg (dict): The config dict for norm layers.
            Defaults to ``dict(type='LN2d', eps=1e-6)``.
        act_cfg (dict): The config dict for activation between pointwise
            convolution. Defaults to ``dict(type='GELU')``.
        in_patch_size (int): The patch size of/? input image patch embedding.
            Defaults to 7.
        in_stride (int): The stride of input image patch embedding.
            Defaults to 4.
        in_pad (int): The padding of input image patch embedding.
            Defaults to 2.
        down_patch_size (int): The patch size of downsampling patch embedding.
            Defaults to 3.
        down_stride (int): The stride of downsampling patch embedding.
            Defaults to 2.
        down_pad (int): The padding of downsampling patch embedding.
            Defaults to 1.
        drop_rate (float): Dropout rate. Defaults to 0.
        drop_path_rate (float): Stochastic depth rate. Defaults to 0.
        out_indices (Sequence | int): Output from which network position.
            Index 0-6 respectively corresponds to
            [stage1, downsampling, stage2, downsampling, stage3, downsampling, stage4]
            Defaults to -1, means the last stage.
        frozen_stages (int): Stages to be frozen (all param fixed).
            Defaults to -1, which means not freezing any parameters.
        deploy (bool): Whether to switch the model structure to
            deployment mode. Default: False.
        init_cfg (dict, optional): Initialization config dict
    """  # noqa: E501

    # --layers: [x,x,x,x], numbers of layers for the four stages
    # --embed_dims, --mlp_ratios:
    #     embedding dims and mlp ratios for the four stages
    # --downsamples: flags to apply downsampling or not in four blocks
    arch_settings = {
        's12': {
            'layers': [2, 2, 6, 2],
            'embed_dims': [64, 128, 320, 512],
            'mlp_ratios': [4, 4, 4, 4],
            'layer_scale_init_value': 1e-5,
        },
        's24': {
            'layers': [4, 4, 12, 4],
            'embed_dims': [64, 128, 320, 512],
            'mlp_ratios': [4, 4, 4, 4],
            'layer_scale_init_value': 1e-5,
        },
        's36': {
            'layers': [6, 6, 18, 6],
            'embed_dims': [64, 128, 320, 512],
            'mlp_ratios': [4, 4, 4, 4],
            'layer_scale_init_value': 1e-6,
        },
        'm36': {
            'layers': [6, 6, 18, 6],
            'embed_dims': [96, 192, 384, 768],
            'mlp_ratios': [4, 4, 4, 4],
            'layer_scale_init_value': 1e-6,
        },
        'm48': {
            'layers': [8, 8, 24, 8],
            'embed_dims': [96, 192, 384, 768],
            'mlp_ratios': [4, 4, 4, 4],
            'layer_scale_init_value': 1e-6,
        },
    }

    def __init__(self,
                 arch='s12',
                 weights = '',
                 in_channels=3,
                 norm_cfg=dict(type='GN', num_groups=1),
                 act_cfg=dict(type='GELU'),
                 in_patch_size=7,
                 in_stride=4,
                 in_pad=2,
                 down_patch_size=3,
                 down_stride=2,
                 down_pad=1,
                 drop_rate=0.,
                 drop_path_rate=0.,
                 out_indices=[0, 2, 4, 6],
                 deploy=False):

        super().__init__()

        if isinstance(arch, str):
            assert arch in self.arch_settings, \
                f'Unavailable arch, please choose from ' \
                f'({set(self.arch_settings)}) or pass a dict.'
            arch = self.arch_settings[arch]
        elif isinstance(arch, dict):
            assert 'layers' in arch and 'embed_dims' in arch, \
                f'The arch dict must have "layers" and "embed_dims", ' \
                f'but got {list(arch.keys())}.'

        layers = arch['layers']
        embed_dims = arch['embed_dims']
        mlp_ratios = arch['mlp_ratios'] \
            if 'mlp_ratios' in arch else [4, 4, 4, 4]
        layer_scale_init_value = arch['layer_scale_init_value'] \
            if 'layer_scale_init_value' in arch else 1e-5

        self.patch_embed = PatchEmbed(
            patch_size=in_patch_size,
            stride=in_stride,
            padding=in_pad,
            in_chans=in_channels,
            embed_dim=embed_dims[0])

        # set the main block in network
        network = []
        for i in range(len(layers)):
            stage = basic_blocks(
                embed_dims[i],
                i,
                layers,
                mlp_ratio=mlp_ratios[i],
                norm_cfg=norm_cfg,
                act_cfg=act_cfg,
                drop_rate=drop_rate,
                drop_path_rate=drop_path_rate,
                layer_scale_init_value=layer_scale_init_value,
                deploy=deploy)
            network.append(stage)
            if i >= len(layers) - 1:
                break
            if embed_dims[i] != embed_dims[i + 1]:
                # downsampling between two stages
                network.append(
                    PatchEmbed(
                        patch_size=down_patch_size,
                        stride=down_stride,
                        padding=down_pad,
                        in_chans=embed_dims[i],
                        embed_dim=embed_dims[i + 1]))

        self.network = nn.ModuleList(network)

        if isinstance(out_indices, int):
            out_indices = [out_indices]
        assert isinstance(out_indices, Sequence), \
            f'"out_indices" must by a sequence or int, ' \
            f'get {type(out_indices)} instead.'
        for i, index in enumerate(out_indices):
            if index < 0:
                out_indices[i] = 7 + index
                assert out_indices[i] >= 0, f'Invalid out_indices {index}'
        self.out_indices = out_indices
        if self.out_indices:
            for i_layer in self.out_indices:
                layer = build_norm_layer(norm_cfg,
                                         embed_dims[(i_layer + 1) // 2])[1]
                layer_name = f'norm{i_layer}'
                self.add_module(layer_name, layer)

        self.deploy = deploy
        if weights:
            self.load_state_dict(update_weight(self.state_dict(), torch.load(weights)['state_dict']))
        self.channel = [i.size(1) for i in self.forward(torch.randn(1, 3, 640, 640))]

    def forward_embeddings(self, x):
        x = self.patch_embed(x)
        return x

    def forward_tokens(self, x):
        outs = []
        for idx, block in enumerate(self.network):
            x = block(x)
            if idx in self.out_indices:
                norm_layer = getattr(self, f'norm{idx}')
                x_out = norm_layer(x)
                outs.append(x_out)
        return outs
    
    def forward(self, x):
        # input embedding
        x = self.forward_embeddings(x)
        # through backbone
        x = self.forward_tokens(x)
        return x

if __name__ == '__main__':
    model = RIFormer('s12', 'riformer-s12_32xb128_in1k-384px_20230406-145eda4c.pth')
    inputs = torch.randn((1, 3, 640, 640))
    for i in model(inputs):
        print(i.size())
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446

Backbone替换

yolo.py修改

def parse_model函数

def parse_model(d, ch):  # model_dict, input_channels(3)
    # Parse a YOLOv5 model.yaml dictionary
    LOGGER.info(f"\n{'':>3}{'from':>18}{'n':>3}{'params':>10}  {'module':<40}{'arguments':<30}")
    anchors, nc, gd, gw, act = d['anchors'], d['nc'], d['depth_multiple'], d['width_multiple'], d.get('activation')
    if act:
        Conv.default_act = eval(act)  # redefine default activation, i.e. Conv.default_act = nn.SiLU()
        LOGGER.info(f"{colorstr('activation:')} {act}")  # print
    na = (len(anchors[0]) // 2) if isinstance(anchors, list) else anchors  # number of anchors
    no = na * (nc + 5)  # number of outputs = anchors * (classes + 5)

    is_backbone = False
    layers, save, c2 = [], [], ch[-1]  # layers, savelist, ch out
    for i, (f, n, m, args) in enumerate(d['backbone'] + d['head']):  # from, number, module, args
        try:
            t = m
            m = eval(m) if isinstance(m, str) else m  # eval strings
        except:
            pass
        for j, a in enumerate(args):
            with contextlib.suppress(NameError):
                try:
                    args[j] = eval(a) if isinstance(a, str) else a  # eval strings
                except:
                    args[j] = a

        n = n_ = max(round(n * gd), 1) if n > 1 else n  # depth gain
        if m in {
                Conv, GhostConv, Bottleneck, GhostBottleneck, SPP, SPPF, DWConv, MixConv2d, Focus, CrossConv,
                BottleneckCSP, C3, C3TR, C3SPP, C3Ghost, nn.ConvTranspose2d, DWConvTranspose2d, C3x}:
            c1, c2 = ch[f], args[0]
            if c2 != no:  # if not output
                c2 = make_divisible(c2 * gw, 8)

            args = [c1, c2, *args[1:]]
            if m in {BottleneckCSP, C3, C3TR, C3Ghost, C3x}:
                args.insert(2, n)  # number of repeats
                n = 1
        elif m is nn.BatchNorm2d:
            args = [ch[f]]
        elif m is Concat:
            c2 = sum(ch[x] for x in f)
        # TODO: channel, gw, gd
        elif m in {Detect, Segment}:
            args.append([ch[x] for x in f])
            if isinstance(args[1], int):  # number of anchors
                args[1] = [list(range(args[1] * 2))] * len(f)
            if m is Segment:
                args[3] = make_divisible(args[3] * gw, 8)
        elif m is Contract:
            c2 = ch[f] * args[0] ** 2
        elif m is Expand:
            c2 = ch[f] // args[0] ** 2
        elif isinstance(m, str):
            t = m
            m = timm.create_model(m, pretrained=args[0], features_only=True)
            c2 = m.feature_info.channels()
        elif m in {RIFormer}: #添加Backbone
            m = m(*args)
            c2 = m.channel
        else:
            c2 = ch[f]
        if isinstance(c2, list):
            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
        np = sum(x.numel() for x in m_.parameters())  # number params
        m_.i, m_.f, m_.type, m_.np = i + 4 if is_backbone else i, f, t, np  # attach index, 'from' index, type, number params
        LOGGER.info(f'{i:>3}{str(f):>18}{n_:>3}{np:10.0f}  {t:<40}{str(args):<30}')  # 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):
            ch.extend(c2)
            for _ in range(5 - len(ch)):
                ch.insert(0, 0)
        else:
            ch.append(c2)
    return nn.Sequential(*layers), sorted(save)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82

def _forward_once函数

def _forward_once(self, x, profile=False, visualize=False):
    y, dt = [], []  # outputs
    for m in 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)
            x = x[-1]
        else:
            x = m(x)  # run
            y.append(x if m.i in self.save else None)  # save output
        if visualize:
            feature_visualization(x, m.type, m.i, save_dir=visualize)
    return x
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23

创建.yaml配置文件

# YOLOv5 🚀 by Ultralytics, GPL-3.0 license
# Parameters
nc: 80  # number of classes
depth_multiple: 0.33  # model depth multiple
width_multiple: 0.25  # layer channel multiple
anchors:
  - [10,13, 16,30, 33,23]  # P3/8
  - [30,61, 62,45, 59,119]  # P4/16
  - [116,90, 156,198, 373,326]  # P5/32

# 0-P1/2
# 1-P2/4
# 2-P3/8
# 3-P4/16
# 4-P5/32

# YOLOv5 v6.0 backbone
backbone:
  # [from, number, module, args]
  [[-1, 1, RIFormer, [False]], # 4
   [-1, 1, SPPF, [1024, 5]],  # 5
  ]

# YOLOv5 v6.0 head
head:
  [[-1, 1, Conv, [512, 1, 1]], # 6
   [-1, 1, nn.Upsample, [None, 2, 'nearest']], # 7
   [[-1, 3], 1, Concat, [1]],  # cat backbone P4 8
   [-1, 3, C3, [512, False]],  # 9

   [-1, 1, Conv, [256, 1, 1]], # 10
   [-1, 1, nn.Upsample, [None, 2, 'nearest']], # 11
   [[-1, 2], 1, Concat, [1]],  # cat backbone P3 12
   [-1, 3, C3, [256, False]],  # 13 (P3/8-small)

   [-1, 1, Conv, [256, 3, 2]], # 14
   [[-1, 10], 1, Concat, [1]],  # cat head P4 15
   [-1, 3, C3, [512, False]],  # 16 (P4/16-medium)

   [-1, 1, Conv, [512, 3, 2]], # 17
   [[-1, 5], 1, Concat, [1]],  # cat head P5 18
   [-1, 3, C3, [1024, False]],  # 19 (P5/32-large)

   [[13, 16, 19], 1, Detect, [nc, anchors]],  # Detect(P3, P4, P5)
  ]
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45