yolov5的记录贴

1. 修改激活函数

wandb

        # self.act = nn.Identity() if act is True else (act if isinstance(act, nn.Module) else nn.Identity())
        # self.act = nn.Tanh() if act is True else (act if isinstance(act, nn.Module) else nn.Identity())
        # self.act = nn.Sigmoid() if act is True else (act if isinstance(act, nn.Module) else nn.Identity())
        # self.act = nn.ReLU() if act is True else (act if isinstance(act, nn.Module) else nn.Identity())
        # self.act = nn.LeakyReLU(0.1) if act is True else (act if isinstance(act, nn.Module) else nn.Identity())
        # self.act = nn.Hardswish() if act is True else (act if isinstance(act, nn.Module) else nn.Identity())
        # self.act = nn.SiLU() if act is True else (act if isinstance(act, nn.Module) else nn.Identity())
        # self.act = Mish() if act is True else (act if isinstance(act, nn.Module) else nn.Identity())
        # self.act = FReLU(c2) if act is True else (act if isinstance(act, nn.Module) else nn.Identity())
        # self.act = AconC(c2) if act is True else (act if isinstance(act, nn.Module) else nn.Identity())
        # self.act = MetaAconC(c2) if act is True else (act if isinstance(act, nn.Module) else nn.Identity())
        # self.act = SiLU_beta(c2) if act is True else (act if isinstance(act, nn.Module) else nn.Identity())
        # self.act = FReLU_noBN_biasFalse(c2) if act is True else (act if isinstance(act, nn.Module) else nn.Identity())
        # self.act = FReLU_noBN_biasTrue(c2) if act is True else (act if isinstance(act, nn.Module) else nn.Identity())

对比实验图：
其中FRelu达到的最好的效果，但是针对与yolov5s小模型比较友好，会增加计算量

2. 修改优化器

其中AdamW再bs为16的情况下表现良好

3.使用deconv或dw deconv

3.2 使用deconv

GitHub

实验图，最高的是16*16的dw deconv

# YOLOv5 v6.0 head
head:
  [[-1, 1, Conv, [512, 1, 1]],
   [-1, 1, nn.Upsample, [None, 2, 'nearest']], # the origin nn.Upsample layer
   # [-1, 1, nn.ConvTranspose2d, [512, 2, 2]], # nn.ConvTransposed2d with 2x2 deconv kernel
   # [-1, 1, nn.ConvTranspose2d, [512, 4, 2, 1]], # nn.ConvTransposed2d with 4x4 deconv kernel
   [[-1, 6], 1, Concat, [1]],  # cat backbone P4
   [-1, 3, C3, [512, False]],  # 13

   [-1, 1, Conv, [256, 1, 1]],
   [-1, 1, nn.Upsample, [None, 2, 'nearest']], # the origin nn.Upsample layer
   # [-1, 1, nn.ConvTranspose2d, [256, 2, 2]], # nn.ConvTransposed2d with 2x2 deconv kernel
   # [-1, 1, nn.ConvTranspose2d, [256, 4, 2, 1]], # nn.ConvTransposed2d with 4x4 deconv kernel
   [[-1, 4], 1, Concat, [1]],  # cat backbone P3
   [-1, 3, C3, [256, False]],  # 17 (P3/8-small)

   [-1, 1, Conv, [256, 3, 2]],
   [[-1, 14], 1, Concat, [1]],  # cat head P4
   [-1, 3, C3, [512, False]],  # 20 (P4/16-medium)

   [-1, 1, Conv, [512, 3, 2]],
   [[-1, 10], 1, Concat, [1]],  # cat head P5
   [-1, 3, C3, [1024, False]],  # 23 (P5/32-large)

   [[17, 20, 23], 1, Detect, [nc, anchors]],  # Detect(P3, P4, P5)
  ]
# the parse_model of yolo.py should be changed from
# if m in [Conv, GhostConv, Bottleneck, GhostBottleneck, SPP, SPPF, DWConv, MixConv2d, Focus, CrossConv,
#         BottleneckCSP, C3, C3TR, C3SPP, C3Ghost]:
# to
if m in [Conv, GhostConv, Bottleneck, GhostBottleneck, SPP, SPPF, DWConv, MixConv2d, Focus, CrossConv,
         BottleneckCSP, C3, C3TR, C3SPP, C3Ghost, nn.ConvTranspose2d]:

2 使用 4*4 dw deconv

# YOLOv5 v6.0 head with Deconvolution Upsample
head:
  [[-1, 1, Conv, [512, 1, 1]],
   [-1, 1, nn.ConvTranspose2d, [512, 4, 2, 1, 0, 512]],
   [[-1, 6], 1, Concat, [1]],  # cat backbone P4
   [-1, 3, C3, [512, False]],  # 13

   [-1, 1, Conv, [256, 1, 1]],
   [-1, 1, nn.ConvTranspose2d, [256, 4, 2, 1, 0, 256]],
   [[-1, 4], 1, Concat, [1]],  # cat backbone P3
   [-1, 3, C3, [256, False]],  # 17 (P3/8-small)

   [-1, 1, Conv, [256, 3, 2]],
   [[-1, 14], 1, Concat, [1]],  # cat head P4
   [-1, 3, C3, [512, False]],  # 20 (P4/16-medium)

   [-1, 1, Conv, [512, 3, 2]],
   [[-1, 10], 1, Concat, [1]],  # cat head P5
   [-1, 3, C3, [1024, False]],  # 23 (P5/32-large)

   [[17, 20, 23], 1, Detect, [nc, anchors]],  # Detect(P3, P4, P5)
  ]
        if m in (Conv, GhostConv, Bottleneck, GhostBottleneck, SPP, SPPF, DWConv, MixConv2d, Focus, CrossConv, BottleneckCSP, C3, C3TR, C3SPP, C3Ghost, nn.ConvTranspose2d):
            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]:
                args.insert(2, n)  # number of repeats
                n = 1
            elif m is nn.ConvTranspose2d:
                if len(args) >= 7:
                    args[6] = make_divisible(args[6] * gw, 8)

3.3 添加dwconv

结论 ：不同位置的dwconv可以在yolov5s上得到涨点，

github
wandb

添加方式：yolov5s-dw5l21，yolov5s代表模型网络结构，dw5l为卷积核为5，21代表为在第21层添加

# YOLOv5  by Ultralytics, GPL-3.0 license

# Parameters
nc: 80  # number of classes
depth_multiple: 0.33  # model depth multiple
width_multiple: 0.50  # 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

# YOLOv5 v6.0 backbone
backbone:
  # [from, number, module, args]
  [[-1, 1, Conv, [64, 6, 2, 2]],  # 0-P1/2
   [-1, 1, Conv, [128, 3, 2]],  # 1-P2/4
   [-1, 3, C3, [128]],
   [-1, 1, Conv, [256, 3, 2]],  # 3-P3/8
   [-1, 6, C3, [256]],
   [-1, 1, Conv, [512, 3, 2]],  # 5-P4/16
   [-1, 9, C3, [512]],
   [-1, 1, DWConv, [1024, 7, 2]],  # 7-P5/32
   [-1, 3, C3, [1024]],
   [-1, 1, SPPF, [1024, 5]],  # 9
  ]

# YOLOv5 v6.0 head
head:
  [[-1, 1, Conv, [512, 1, 1]],
   [-1, 1, nn.Upsample, [None, 2, 'nearest']],
   [[-1, 6], 1, Concat, [1]],  # cat backbone P4
   [-1, 3, C3, [512, False]],  # 13

   [-1, 1, Conv, [256, 1, 1]],
   [-1, 1, nn.Upsample, [None, 2, 'nearest']],
   [[-1, 4], 1, Concat, [1]],  # cat backbone P3
   [-1, 3, C3, [256, False]],  # 17 (P3/8-small)

   [-1, 1, Conv, [256, 3, 2]],
   [[-1, 14], 1, Concat, [1]],  # cat head P4
   [-1, 3, C3, [512, False]],  # 20 (P4/16-medium)

   [-1, 1, DWConv, [512, 7, 2]],
   [[-1, 10], 1, Concat, [1]],  # cat head P5
   [-1, 3, C3, [1024, False]],  # 23 (P5/32-large)

   [[17, 20, 23], 1, Detect, [nc, anchors]],  # Detect(P3, P4, P5)
  ]

other

obj、cls、loss解释

训练拟合结果分析

智星云平台租借gpu