Dear author,
Many thanks for your great work.
Just found an issue with train.py:
bxy = [[yb // length_0, xb // length_1] for (yb, xb) in coord]
will give you 0 values in the target building process. Could it be a minor mistake, or caused by a different version of python?
How about replacing it with:
bxy= [[torch.true_divide(yb, length_0),torch.true_divide(xb, length_1)] for (yb, xb) in coord]
Dear author,
Many thanks for your great work.
Just found an issue with train.py:
bxy = [[yb // length_0, xb // length_1] for (yb, xb) in coord]
will give you 0 values in the target building process. Could it be a minor mistake, or caused by a different version of python?
How about replacing it with:
bxy= [[torch.true_divide(yb, length_0),torch.true_divide(xb, length_1)] for (yb, xb) in coord]