Releases: gschramm/parallelproj
Releases · gschramm/parallelproj
v1.6.0
Changelog vs v1.5.0
Changes to the LinearOperator abstract base class break backward compatibility
C/CUDA libraries + API were not changed
Changes to the python API:
- rewrite LinearOperator base class + basic operators to support python array api including different devices
- add missing type hints
- add finite difference operator
- remove obsolete functions
v1.5
Changelog vs v1.4
- add compatibility of python wrapper to python array api (via array-api-compat) such that numpy, cupy, pytorch arrays can be directly projected using the python wrappers
- a few small breaking changes in the python wrappers to the projectors, but no changes to the C/CUDA libs
- update of the python tests and examples