search

WongKinYiu / yolov9

Implementation of paper - YOLOv9: Learning What You Want to Learn Using Programmable Gradient Information
GNU General Public License v3.0
8.63k stars 1.31k forks source link

How to use TripleDDetect? #226

Open huihui308 opened 4 months ago

huihui308 commented 4 months ago

There is a TripleDDetect in yolo.py, how to use it? Could you give me a hint?

Youho99 commented 4 months ago

I think you should use train_triple.py

huihui308 commented 4 months ago

I think you should use train_triple.py

Yes, you are right, but I think it needs a model config file to fit triple head.

Youho99 commented 4 months ago

Have you tried ?

huihui308 commented 4 months ago

Have you tried ?

# YOLOv9

# parameters
nc: 7  # number of classes
depth_multiple: 1.0  # model depth multiple
width_multiple: 1.0  # layer channel multiple
#activation: nn.LeakyReLU(0.1)
#activation: nn.ReLU()

# anchors
anchors: 3

# YOLOv9 backbone
backbone:
  [
   [-1, 1, Silence, []],  

   # conv down
   [-1, 1, Conv, [64, 3, 2]],  # 1-P1/2

   # conv down
   [-1, 1, Conv, [128, 3, 2]],  # 2-P2/4

   # elan-1 block
   [-1, 1, RepNCSPELAN4, [256, 128, 64, 1]],  # 3

   # avg-conv down
   [-1, 1, ADown, [256]],  # 4-P3/8

   # elan-2 block
   [-1, 1, RepNCSPELAN4, [512, 256, 128, 1]],  # 5

   # avg-conv down
   [-1, 1, ADown, [512]],  # 6-P4/16

   # elan-2 block
   [-1, 1, RepNCSPELAN4, [512, 512, 256, 1]],  # 7

   # avg-conv down
   [-1, 1, ADown, [512]],  # 8-P5/32

   # elan-2 block
   [-1, 1, RepNCSPELAN4, [512, 512, 256, 1]],  # 9
  ]

# YOLOv9 head
head:
  [
   # elan-spp block
   [-1, 1, SPPELAN, [512, 256]],  # 10

   # up-concat merge
   [-1, 1, nn.Upsample, [None, 2, 'nearest']],
   [[-1, 7], 1, Concat, [1]],  # cat backbone P4

   # elan-2 block
   [-1, 1, RepNCSPELAN4, [512, 512, 256, 1]],  # 13

   # up-concat merge
   [-1, 1, nn.Upsample, [None, 2, 'nearest']],
   [[-1, 5], 1, Concat, [1]],  # cat backbone P3

   # elan-2 block
   [-1, 1, RepNCSPELAN4, [256, 256, 128, 1]],  # 16 (P3/8-small)

   # avg-conv-down merge
   [-1, 1, ADown, [256]],
   [[-1, 13], 1, Concat, [1]],  # cat head P4

   # elan-2 block
   [-1, 1, RepNCSPELAN4, [512, 512, 256, 1]],  # 19 (P4/16-medium)

   # avg-conv-down merge
   [-1, 1, ADown, [512]],
   [[-1, 10], 1, Concat, [1]],  # cat head P5

   # elan-2 block
   [-1, 1, RepNCSPELAN4, [512, 512, 256, 1]],  # 22 (P5/32-large)

   # multi-level reversible auxiliary branch

   # routing
   [5, 1, CBLinear, [[256]]], # 23
   [7, 1, CBLinear, [[256, 512]]], # 24
   [9, 1, CBLinear, [[256, 512, 512]]], # 25

   # conv down
   [0, 1, Conv, [64, 3, 2]],  # 26-P1/2

   # conv down
   [-1, 1, Conv, [128, 3, 2]],  # 27-P2/4

   # elan-1 block
   [-1, 1, RepNCSPELAN4, [256, 128, 64, 1]],  # 28

   # avg-conv down fuse
   [-1, 1, ADown, [256]],  # 29-P3/8
   [[23, 24, 25, -1], 1, CBFuse, [[0, 0, 0]]], # 30  

   # elan-2 block
   [-1, 1, RepNCSPELAN4, [512, 256, 128, 1]],  # 31

   # avg-conv down fuse
   [-1, 1, ADown, [512]],  # 32-P4/16
   [[24, 25, -1], 1, CBFuse, [[1, 1]]], # 33 

   # elan-2 block
   [-1, 1, RepNCSPELAN4, [512, 512, 256, 1]],  # 34

   # avg-conv down fuse
   [-1, 1, ADown, [512]],  # 35-P5/32
   [[25, -1], 1, CBFuse, [[2]]], # 36

   # elan-2 block
   [-1, 1, RepNCSPELAN4, [512, 512, 256, 1]],  # 37

   # routing
   [5, 1, CBLinear, [[256]]], # 38
   [7, 1, CBLinear, [[256, 512]]], # 39
   [9, 1, CBLinear, [[256, 512, 512]]], # 40

   # conv down
   [0, 1, Conv, [64, 3, 2]],  # 41-P1/2

   # conv down
   [-1, 1, Conv, [128, 3, 2]],  # 42-P2/4

   # elan-1 block
   [-1, 1, RepNCSPELAN4, [256, 128, 64, 1]],  # 43

   # avg-conv down fuse
   [-1, 1, ADown, [256]],  # 44-P3/8
   [[23, 24, 25, -1], 1, CBFuse, [[0, 0, 0]]], # 45  

   # elan-2 block
   [-1, 1, RepNCSPELAN4, [512, 256, 128, 1]],  # 46

   # avg-conv down fuse
   [-1, 1, ADown, [512]],  # 47-P4/16
   [[24, 25, -1], 1, CBFuse, [[1, 1]]], # 48 

   # elan-2 block
   [-1, 1, RepNCSPELAN4, [512, 512, 256, 1]],  # 49

   # avg-conv down fuse
   [-1, 1, ADown, [512]],  # 50-P5/32
   [[25, -1], 1, CBFuse, [[2]]], # 51

   # elan-2 block
   [-1, 1, RepNCSPELAN4, [512, 512, 256, 1]],  # 52

   # detection head

   # detect
   [[31, 34, 37, 46, 49, 52, 16, 19, 22], 1, TripleDetect, [nc]],  # TripleDetect(A3, A4, A5, B3, B4, B5, P3, P4, P5)
  ]
Nomiluks commented 3 months ago

What do mean by values like A3, A4, A5, B3, B4, B5, P3, P4, P5 ?

tigerzjh commented 3 months ago

Have you tried ?

# YOLOv9

# parameters
nc: 7  # number of classes
depth_multiple: 1.0  # model depth multiple
width_multiple: 1.0  # layer channel multiple
#activation: nn.LeakyReLU(0.1)
#activation: nn.ReLU()

# anchors
anchors: 3

# YOLOv9 backbone
backbone:
  [
   [-1, 1, Silence, []],  

   # conv down
   [-1, 1, Conv, [64, 3, 2]],  # 1-P1/2

   # conv down
   [-1, 1, Conv, [128, 3, 2]],  # 2-P2/4

   # elan-1 block
   [-1, 1, RepNCSPELAN4, [256, 128, 64, 1]],  # 3

   # avg-conv down
   [-1, 1, ADown, [256]],  # 4-P3/8

   # elan-2 block
   [-1, 1, RepNCSPELAN4, [512, 256, 128, 1]],  # 5

   # avg-conv down
   [-1, 1, ADown, [512]],  # 6-P4/16

   # elan-2 block
   [-1, 1, RepNCSPELAN4, [512, 512, 256, 1]],  # 7

   # avg-conv down
   [-1, 1, ADown, [512]],  # 8-P5/32

   # elan-2 block
   [-1, 1, RepNCSPELAN4, [512, 512, 256, 1]],  # 9
  ]

# YOLOv9 head
head:
  [
   # elan-spp block
   [-1, 1, SPPELAN, [512, 256]],  # 10

   # up-concat merge
   [-1, 1, nn.Upsample, [None, 2, 'nearest']],
   [[-1, 7], 1, Concat, [1]],  # cat backbone P4

   # elan-2 block
   [-1, 1, RepNCSPELAN4, [512, 512, 256, 1]],  # 13

   # up-concat merge
   [-1, 1, nn.Upsample, [None, 2, 'nearest']],
   [[-1, 5], 1, Concat, [1]],  # cat backbone P3

   # elan-2 block
   [-1, 1, RepNCSPELAN4, [256, 256, 128, 1]],  # 16 (P3/8-small)

   # avg-conv-down merge
   [-1, 1, ADown, [256]],
   [[-1, 13], 1, Concat, [1]],  # cat head P4

   # elan-2 block
   [-1, 1, RepNCSPELAN4, [512, 512, 256, 1]],  # 19 (P4/16-medium)

   # avg-conv-down merge
   [-1, 1, ADown, [512]],
   [[-1, 10], 1, Concat, [1]],  # cat head P5

   # elan-2 block
   [-1, 1, RepNCSPELAN4, [512, 512, 256, 1]],  # 22 (P5/32-large)

   # multi-level reversible auxiliary branch

   # routing
   [5, 1, CBLinear, [[256]]], # 23
   [7, 1, CBLinear, [[256, 512]]], # 24
   [9, 1, CBLinear, [[256, 512, 512]]], # 25

   # conv down
   [0, 1, Conv, [64, 3, 2]],  # 26-P1/2

   # conv down
   [-1, 1, Conv, [128, 3, 2]],  # 27-P2/4

   # elan-1 block
   [-1, 1, RepNCSPELAN4, [256, 128, 64, 1]],  # 28

   # avg-conv down fuse
   [-1, 1, ADown, [256]],  # 29-P3/8
   [[23, 24, 25, -1], 1, CBFuse, [[0, 0, 0]]], # 30  

   # elan-2 block
   [-1, 1, RepNCSPELAN4, [512, 256, 128, 1]],  # 31

   # avg-conv down fuse
   [-1, 1, ADown, [512]],  # 32-P4/16
   [[24, 25, -1], 1, CBFuse, [[1, 1]]], # 33 

   # elan-2 block
   [-1, 1, RepNCSPELAN4, [512, 512, 256, 1]],  # 34

   # avg-conv down fuse
   [-1, 1, ADown, [512]],  # 35-P5/32
   [[25, -1], 1, CBFuse, [[2]]], # 36

   # elan-2 block
   [-1, 1, RepNCSPELAN4, [512, 512, 256, 1]],  # 37

   # routing
   [5, 1, CBLinear, [[256]]], # 38
   [7, 1, CBLinear, [[256, 512]]], # 39
   [9, 1, CBLinear, [[256, 512, 512]]], # 40

   # conv down
   [0, 1, Conv, [64, 3, 2]],  # 41-P1/2

   # conv down
   [-1, 1, Conv, [128, 3, 2]],  # 42-P2/4

   # elan-1 block
   [-1, 1, RepNCSPELAN4, [256, 128, 64, 1]],  # 43

   # avg-conv down fuse
   [-1, 1, ADown, [256]],  # 44-P3/8
   [[23, 24, 25, -1], 1, CBFuse, [[0, 0, 0]]], # 45  

   # elan-2 block
   [-1, 1, RepNCSPELAN4, [512, 256, 128, 1]],  # 46

   # avg-conv down fuse
   [-1, 1, ADown, [512]],  # 47-P4/16
   [[24, 25, -1], 1, CBFuse, [[1, 1]]], # 48 

   # elan-2 block
   [-1, 1, RepNCSPELAN4, [512, 512, 256, 1]],  # 49

   # avg-conv down fuse
   [-1, 1, ADown, [512]],  # 50-P5/32
   [[25, -1], 1, CBFuse, [[2]]], # 51

   # elan-2 block
   [-1, 1, RepNCSPELAN4, [512, 512, 256, 1]],  # 52

   # detection head

   # detect
   [[31, 34, 37, 46, 49, 52, 16, 19, 22], 1, TripleDetect, [nc]],  # TripleDetect(A3, A4, A5, B3, B4, B5, P3, P4, P5)
  ]

where is the yaml, i search the repo，there is no config use tripleDDetect

tigerzjh commented 3 months ago

where is the yaml, i search the repo，there is no config use tripleDDetect

where is the yaml, i search the repo，there is no config use tripleDDetect。 image

Youho99 commented 3 months ago

@tigerzjh Read https://github.com/WongKinYiu/yolov9/issues/1#issuecomment-1958742287 Triple is maybe a template ?