84d32df24d9b88464ed851dadda9bdd74ad67c5a,tests/test_retina.py,,denormalize,#,19

Before Change

    is the size of the image.
    
    x = 0.5 * ((coords[:, 0] + 1.0) * T)
    y = 0.5 * ((coords[:, 1] + 1.0) * T)
    return torch.stack([x, y], dim=1).long()

After Change

    coordinates in the range [0, T] where T is
    the size of the image.
    
    return (0.5 * ((coords + 1.0) * T)).long()


def bounding_box(x, y, size, color="w"):
    x = int(x - (size / 2))

Instances