Hi,
Thank you for the great work. I really enjoyed the paper particularly about how the discrete variables are treated.
I have a minor question regarding to the losses about translation and rotations. In the paper, the denominator is (1-t), where as in the code, it seems to be (1-t)^2, see the screenshot below.
My understanding is that (1-t)^2 in the code should be correct. Could you help double confirm it?
Thanks,
Ruijiang
Hi,
Thank you for the great work. I really enjoyed the paper particularly about how the discrete variables are treated.
I have a minor question regarding to the losses about translation and rotations. In the paper, the denominator is (1-t), where as in the code, it seems to be (1-t)^2, see the screenshot below.
My understanding is that (1-t)^2 in the code should be correct. Could you help double confirm it?
Thanks,
Ruijiang