You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
This is an issue because we calculate the output shape as out_shape = in_shape * scale + offset * 2, so we need to allow for fractional output values in case the target value for offset * 2 is an odd number.
The text was updated successfully, but these errors were encountered:
related question: for such uneven case, how would one determine the relative position of output wrt input?
(we might want to ignore this one for now and just leave it to the judgement of any consumer software)
related question: for such uneven case, how would one determine the relative position of output wrt input?
(we might want to ignore this one for now and just leave it to the judgement of any consumer software)
I think this case mostly comes up for the channel axes in order to specify the number of output channels as scale=0, offset=n_channels/2. So I would ignore this issue for now.
This is an issue because we calculate the output shape as
out_shape = in_shape * scale + offset * 2
, so we need to allow for fractional output values in case the target value foroffset * 2
is an odd number.The text was updated successfully, but these errors were encountered: