Can I train 4-channel pictures?

[LVROBOT]

Did some data processing，Can I train 4-channel pictures?

[glenn-jocher]

@LVROBOT sure, you can train any channel images. When models are created you can specify how many input channels are used, though beware opencv will typically load most images, even black-white images as 3-ch images.

The Model() accepts a channel argument here, with ch=3 by default.

yolov5/models/yolo.py

Lines 45 to 46 in 121d90b

	class Model(nn.Module):
	def __init__(self, model_cfg='yolov5s.yaml', ch=3, nc=None): # model, input channels, number of classes

Also when calling a model from torch hub you can set the input channel count similarly. See https://docs.ultralytics.com/yolov5/tutorials/pytorch_hub_model_loading

[LVROBOT]

@glenn-jocher
I train 4-channel pictures, error is :

0%| | 0/1196 [00:07<?, ?it/s]
Traceback (most recent call last):
File "/home/leo/Desktop/DL/yolov5-master/train.py", line 409, in
train(hyp)
File "/home/leo/Desktop/DL/yolov5-master/train.py", line 258, in train
pred = model(imgs)
File "/home/leo/anaconda3/envs/torch/lib/python3.6/site-packages/torch/nn/modules/module.py", line 532, in call
result = self.forward(*input, **kwargs)
File "/home/leo/Desktop/DL/yolov5-master/models/yolo.py", line 90, in forward
return self.forward_once(x, profile) # single-scale inference, train
File "/home/leo/Desktop/DL/yolov5-master/models/yolo.py", line 107, in forward_once
x = m(x) # run
File "/home/leo/anaconda3/envs/torch/lib/python3.6/site-packages/torch/nn/modules/module.py", line 532, in call
result = self.forward(*input, **kwargs)
File "/home/leo/Desktop/DL/yolov5-master/models/common.py", line 86, in forward
return self.conv(torch.cat([x[..., ::2, ::2], x[..., 1::2, ::2], x[..., ::2, 1::2], x[..., 1::2, 1::2]], 1))
File "/home/leo/anaconda3/envs/torch/lib/python3.6/site-packages/torch/nn/modules/module.py", line 532, in call
result = self.forward(*input, **kwargs)
File "/home/leo/Desktop/DL/yolov5-master/models/common.py", line 21, in forward
return self.act(self.bn(self.conv(x)))
File "/home/leo/anaconda3/envs/torch/lib/python3.6/site-packages/torch/nn/modules/module.py", line 532, in call
result = self.forward(*input, **kwargs)
File "/home/leo/anaconda3/envs/torch/lib/python3.6/site-packages/torch/nn/modules/conv.py", line 345, in forward
return self.conv2d_forward(input, self.weight)
File "/home/leo/anaconda3/envs/torch/lib/python3.6/site-packages/torch/nn/modules/conv.py", line 342, in conv2d_forward
self.padding, self.dilation, self.groups)
RuntimeError: Given groups=1, weight of size 32 16 3 3, expected input[8, 12, 320, 320] to have 16 channels, but got 12 channels instead

[github-actions]

This issue has been automatically marked as stale because it has not had recent activity. It will be closed if no further activity occurs. Thank you for your contributions.