If anybody could send me just one example of torch.tensor for multicalss detection training I would be very thankfull.
From pytorch tutorial I found only oneclass detection dataset.
But if I have several classes:
What should be the box for image if there is no object with this label on it?
Probably it should be [0,0,0,0] but I’m not sure…
Correct me if I’m not right, labels should be one-hot-encoded for multiclass detection?
And I don’t understand how boxes tensor should look like if there are several objects for 1 label on the picture?
Please help, pytorch tutorial is completely useless