model.fuse() reports error messag

[dongzhenguo2016]

Originally model.fuse() in detect.py was commented. I want to release this line of code to see if merging conv with bn really improves the speed of inference, but the code reports the following error message:

Using CUDA device0 CudaDeviceProperties(name='GeForce RTX 2080 Ti', total_memory=10989MB)
Model Summary: 225 layers, 6.26003e+07 parameters, 6.26003e+07 gradients
Fusing layers...
Traceback (most recent call last):
File "/media/hdc/data1/dzc/yolov3-master/detect.py", line 223, in
detect()
File "/media/hdc/data1/dzc/yolov3-master/detect.py", line 61, in detect
model.fuse()
File "/media/hdc/data1/dzc/yolov3-master/models.py", line 334, in fuse
fused = torch_utils.fuse_conv_and_bn(conv, b)
File "/media/hdc/data1/dzc/yolov3-master/utils/torch_utils.py", line 93, in fuse_conv_and_bn
fusedconv.bias.copy(torch.mm(w_bn, b_conv.reshape(-1, 1)).reshape(-1) + b_bn)# .cuda(device)
RuntimeError: Expected object of device type cuda but got device type cpu for argument #2 'mat2' in call to _th_mm

Process finished with exit code 1

I tried to search for some solutions, but none of them solved. How can I fix this bug?
thanks a lot

[glenn-jocher]

@dongzhenguo2016 your code is way out of date. Update.