0.8.0
- Major updates to PSF modeling to massively reduce reconstruction times
- Option to have
torch.float16
as default dtype to reduce memory usage in half for those with smaller graphics cards (and possibly speed up reconstruction) - Many many more small updates