Releases: grimme-lab/dxtb
Releases · grimme-lab/dxtb
Release list
v0.4.0
Tested up to Python 3.14 and PyTorch 2.10.0.
What's Changed
- PyTorch 2.9.0 by @marvinfriede in #245
- Update GH actions by @marvinfriede in #251
- Dependency update by @marvinfriede in #252
- Python 3.13 and 3.14 by @marvinfriede in #253
- Bump version to 0.4.0 by @marvinfriede in #254
Full Changelog: v0.3.0...v0.4.0
v0.3.0
What's Changed
- Pin dependencies and fix docs by @marvinfriede in #232
- Dependency update by @marvinfriede in #236
Full Changelog: v0.2.0...v0.3.0
v0.2.0
GFN2-xTB is now available. All GFN parameters are wrapped in torch.nn.Parameter.
What's Changed
- Pytorch 2.5.0 by @marvinfriede in #181
- Improve deletion of timer by @marvinfriede in #176
- Update workflows (PyTorch 2.5.1, zizmor security fixes) by @marvinfriede in #188
- Profile moving devices by @marvinfriede in #190
- GFN2 Core Hamiltonian by @marvinfriede in #191
- GFN2: Self-consistent DFT-D4 by @marvinfriede in #193
- Fix device in full SCF by @marvinfriede in #196
- GFN2: Shell-resolved 3rd order ES by @marvinfriede in #197
- Bump artifact actions to V4 by @marvinfriede in #199
- Improve docs by @marvinfriede in #201
- GFN2: Non-self-consistent D4 ATM term by @marvinfriede in #200
- Fix
tad-mctcAPI calls by @marvinfriede in #203 - GFN2: Basis set by @marvinfriede in #202
- GFN2: Multipole Electrostatics (AES2) by @marvinfriede in #204
- GFN2: Tests by @marvinfriede in #207
- Disable timer by default by @marvinfriede in #208
- Pydantic tensor serialization by @marvinfriede in #211
- Dynamic damping by @marvinfriede in #212
- Fix CUDA tests and consistent signatures by @marvinfriede in #215
- Differentiable parametrization by @marvinfriede in #216
- Fix exported basis format by @marvinfriede in #217
- Make charge and spin always 1D by @marvinfriede in #218
- Fix gradients for basis set by @marvinfriede in #219
- Bug Fix: return correct number of SCF iterations by @ineporozhnii in #222
- Test PyTorch 2.7.0 by @marvinfriede in #224
- Bump version to 0.2.0 by @marvinfriede in #225
- Change default SCF mode by @marvinfriede in #227
New Contributors
- @ineporozhnii made their first contribution in #222
Full Changelog: v0.1.1...v0.2.0
v0.1.1
Small bug fix release.
What's Changed
- Update conda availability by @marvinfriede in #170
- Support PyTorch 2.4.1 by @marvinfriede in #171
- Support Python 3.12 by @marvinfriede in #172
- Add example for challenging SCF convergence by @marvinfriede in #173
- Fix batched charges by @marvinfriede in #180
Full Changelog: v0.1.0...v0.1.1
v0.1.0
Windows and macOS support (without the libcint interface).
What's Changed
- Add citation by @marvinfriede in #160
- Improve code quality by @marvinfriede in #157
- Refactor integral code by @marvinfriede in #162
- Fix code scanning issues by @marvinfriede in #165
- Make
libcintinterface optional by @marvinfriede in #166 - Fix docs and consistency checks in calculator getters by @marvinfriede in #168
Full Changelog: v0.0.1...v0.1.0