I am trying to custom a model with ghostnet base on yolov5

[hawkinglai]

❔Question

This is my model config file base on this paper: GhostNet: More Features from Cheap Operations
Link: https://arxiv.org/abs/1911.11907
I am a beginner of deep learning and thanks for your patient answer

YOLOv5 backbone

backbone:

[from, number, module, args]

[[-1, 1, Focus, [64, 3]], # 0-P1/2
[-1, 1, Conv, [128, 3, 2]], # 1-P2/4
[-1, 3, GhostBottleneck, [128, 3, 1]],
[-1, 1, GhostBottleneck, [256, 3, 2]], # 3-P3/8
[-1, 9, GhostBottleneck, [256, 3, 1]],
[-1, 1, GhostBottleneck, [512, 3, 2]], # 5-P4/16
[-1, 9, GhostBottleneck, [512, 3, 1]],
[-1, 1, Conv, [1024, 3, 2]], # 7-P5/32
[-1, 1, SPP, [1024, [5, 9, 13]]],
]

YOLOv5 head

head:
[[-1, 3, GhostBottleneck, [1024, 3, 1]], #10

[-1, 1, DWConv, [512, 1, 1]],
[-1, 1, nn.Upsample, [None, 2, 'nearest']],
[[-1, 8], 1, Concat, [1]],
[-1, 1, DWConv, [512, 3, 1]],
[-1, 3, GhostBottleneck, [512, 3, 1]], #15

[-1, 1, DWConv, [256, 1, 1]],
[-1, 1, nn.Upsample, [None, 2, 'nearest']],
[[-1, 5], 1, Concat, [1]],
[-1, 1, DWConv, [256, 3, 1]],
[-1, 3, GhostBottleneck, [256, 3, 1]],
[-1, 1, nn.Conv2d, [na * (nc + 5), 1, 1]],

[-2, 1, DWConv, [256, 3, 2]],
[[-1, 15], 1, Concat, [1]],
[-1, 3, GhostBottleneck, [512, 3, 1]],
[-1, 1, nn.Conv2d, [na * (nc + 5), 1, 1]],

[-2, 1, DWConv, [512, 3, 2]],
[[ -1, 10], 1, Concat, [1]],
[-1, 3, GhostBottleneck, [1024, 3, 1]],
[-1, 1, nn.Conv2d, [na * (nc + 5), 1, 1]],

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

Additional context

And after these operations, I ran yolo.py to try this model. And yolo.py report this error:
Sizes of tensors must match except in dimension 1. Got 4 and 2 in dimension 2 (The offending index is 1)

Paste the full text:
from n params module arguments
0 -1 1 3520 models.common.Focus [3, 32, 3]
1 -1 1 18560 models.common.Conv [32, 64, 3, 2]
2 -1 1 3440 models.experimental.GhostBottleneck [64, 64, 3, 1]
3 -1 1 18784 models.experimental.GhostBottleneck [64, 128, 3, 2]
4 -1 3 32928 models.experimental.GhostBottleneck [128, 128, 3, 1]
5 -1 1 66240 models.experimental.GhostBottleneck [128, 256, 3, 2]
6 -1 3 115008 models.experimental.GhostBottleneck [256, 256, 3, 1]
7 -1 1 1180672 models.common.Conv [256, 512, 3, 2]
8 -1 1 656896 models.common.SPP [512, 512, [5, 9, 13]]
9 -1 1 142208 models.experimental.GhostBottleneck [512, 512, 3, 1]
10 -1 1 1024 n DWConv at 0x000001F4DBB19E5 [512, 256, 1, 1]
11 -1 1 0 torch.nn.modules.upsampling.Upsample [None, 2, 'nearest']
12 [-1, 7] 1 0 models.common.Concat [1]
13 -1 1 7424 n DWConv at 0x000001F4DBB19E5 [768, 256, 3, 1]
14 -1 1 38336 models.experimental.GhostBottleneck [256, 256, 3, 1]
15 -1 1 512 n DWConv at 0x000001F4DBB19E5 [256, 128, 1, 1]
16 -1 1 0 torch.nn.modules.upsampling.Upsample [None, 2, 'nearest']
17 [-1, 5] 1 0 models.common.Concat [1]
18 -1 1 3712 n DWConv at 0x000001F4DBB19E5 [384, 128, 3, 1]
19 -1 1 10976 models.experimental.GhostBottleneck [128, 128, 3, 1]
20 -1 1 32895 torch.nn.modules.conv.Conv2d [128, 255, 1, 1]
21 -2 1 1408 n DWConv at 0x000001F4DBB19E5 [128, 128, 3, 2]
22 [-1, 15] 1 0 models.common.Concat [1]
23 -1 1 38336 models.experimental.GhostBottleneck [256, 256, 3, 1]
24 -1 1 65535 torch.nn.modules.conv.Conv2d [256, 255, 1, 1]
25 -2 1 2816 n DWConv at 0x000001F4DBB19E5 [256, 256, 3, 2]
26 [-1, 10] 1 0 models.common.Concat [1]
27 -1 1 142208 models.experimental.GhostBottleneck [512, 512, 3, 1]
28 -1 1 130815 torch.nn.modules.conv.Conv2d [512, 255, 1, 1]
29 [17, 20, 23] 1 228990 Detect [80, [[10, 13, 16, 30, 33, 23], [30, 61, 62, 45, 59, 119], [116, 90, 156, 198, 373, 326]], [384, 255, 256]]
Traceback (most recent call last):
File "D:/yolov/yolov5-comment-master/models/yolo.py", line 251, in
model = Model(opt.cfg).to(device)
File "D:/yolov/yolov5-comment-master/models/yolo.py", line 79, in init
m.stride = torch.tensor([s / x.shape[-2] for x in self.forward(torch.zeros(1, ch, s, s))]) # forward
File "D:/yolov/yolov5-comment-master/models/yolo.py", line 109, in forward
return self.forward_once(x, profile) # single-scale inference, train
File "D:/yolov/yolov5-comment-master/models/yolo.py", line 129, in forward_once
x = m(x) # run
File "D:\anaconda_yolo\envs\yolov5\lib\site-packages\torch\nn\modules\module.py", line 722, in _call_impl
result = self.forward(*input, **kwargs)
File "D:\yolov\yolov5-comment-master\models\common.py", line 100, in forward
return torch.cat(x, self.d)
RuntimeError: Sizes of tensors must match except in dimension 1. Got 8 and 4 in dimension 2 (The offending index is 1)

Process finished with exit code 1

[github-actions]

Hello @KitaharaYukimori, thank you for your interest in our work! Please visit our Custom Training Tutorial to get started, and see our Jupyter Notebook , Docker Image, and Google Cloud Quickstart Guide for example environments.

If this is a bug report, please provide screenshots and minimum viable code to reproduce your issue, otherwise we can not help you.

If this is a custom model or data training question, please note Ultralytics does not provide free personal support. As a leader in vision ML and AI, we do offer professional consulting, from simple expert advice up to delivery of fully customized, end-to-end production solutions for our clients, such as:

• Cloud-based AI systems operating on hundreds of HD video streams in realtime.
• Edge AI integrated into custom iOS and Android apps for realtime 30 FPS video inference.
• Custom data training, hyperparameter evolution, and model exportation to any destination.

For more information please visit https://www.ultralytics.com.

[glenn-jocher]

@KitaharaYukimori there was a GhostConv() bug fix recently, you should clone the latest and see if the error persists. See #1176