OpenMM-Torch 0.5

This release adds the ability to output the computed atomic force directly (#52) without using the back-propagation of energy; and fixes a long-standing issue the CUDA context synchronization (#47, #49), which was causing crashes or incorrect results. All the users are advised to update.

What's Changed

• Improved coordination of CUDA contexts with PyTorch by @peastman in #47
• Update CI by @raimis in #53
• Test if a PyTorch module receives corrects arguments by @raimis in #50
• Models can directly output forces by @peastman in #52
• Fix the synchronisation between the CUDA contexts by @raimis in #49

Full Changelog: v0.4...v0.5

OpenMM-Torch 0.4

This release fixes an error that could cause incorrect force accumulation on the Reference platform.

What's Changed

• Fix the force accumulation in the reference platform by @raimis in #46

Full Changelog: v.03...v0.4

OpenMM-Torch 0.3

This release fixes an error where serializing a TorchForce did not save all information.

What's Changed

• Added missing information to serialization proxy by @peastman in #38

Full Changelog: v0.2...v.03

OpenMM-Torch 0.2

This release fixes an error in the Python API that prevented using global parameters.

What's Changed

• TorchForce supports global parameters by @peastman in #18
• README: Add instructions for conda installation by @jaimergp in #21
• Update README to be more comprehensible by @jchodera in #23
• add CI by @jaimergp in #25
• Added global parameters to Python wrapper by @peastman in #29

New Contributors

• @jaimergp made their first contribution in #21
• @jchodera made their first contribution in #23

Full Changelog: v0.1...v0.2

OpenMM-Torch 0.1

This is the first official release. The code is fully functional and suitable for production use, but the API might change in the future.

What's Changed

• Can run PyTorch calculations on GPU by @peastman in #1
• Removed some obsolete code from build scripts by @peastman in #3
• Added functions for casting by @peastman in #6
• Renamed class and namespace by @peastman in #7
• Created Python test by @peastman in #8
• Fix a race condition by explicitly coping Torch forces by @raimis in #16
• [BUGFIX] Synchronize CUDA context before force copying by @yaoyic in #17
• README describes how to identify OpenMM directory by @peastman in #19

New Contributors

• @yaoyic made their first contribution in #17

Full Changelog: https://github.com/openmm/openmm-torch/commits/v0.1