Closed kulkarnikeerti closed 1 year ago
@kulkarnikeerti You need to make sure that the sizes (width and height) of the input images are all multiples of 32 (such as 640x480), otherwise you will encounter the error you show, since the maximum output stride of the network is 32.
I get this error when
RuntimeError: The size of tensor a (45) must match the size of tensor b (46) at non-singleton dimension 2I
use(720,1440)
as an inputimage size
. But the same model works without any issues when I use(960,1920)
as image size. How does this image sizes differs? Both of them are rectangle.https://github.com/yjh0410/PyTorch_YOLO-Family/blob/a26e9ddd3c2263d6384bee0b0f9ef353775fb281/models/yolo/yolo_nano.py#L204
Can you please tell me where its going wrong. It would be great if you can guide me here. Thanks