Commits · master · Dominik Vilsmeier / DiPAS · GitLab

Feb 03, 2022
- Adjust badges to point to main branch · e22e7fad
  Dominik Vilsmeier authored Feb 03, 2022
  e22e7fad
- Merge branch 'release/v2.0' · c205b8ea
  Dominik Vilsmeier authored Feb 03, 2022
  
  View commits for tag v2.0 v2.0
  
  c205b8ea
- Allow selection of PyTorch backend via either 'pytorch' or 'torch' · 2ed63ce4
  Dominik Vilsmeier authored Feb 03, 2022
  
  2ed63ce4
- Update changelog: include v2.0 · 5d776602
  Dominik Vilsmeier authored Feb 03, 2022
  
  5d776602
- Fix ModuleProxy class to expose 'children' and 'parameters' · a68d69b3
  Dominik Vilsmeier authored Feb 03, 2022
  
  a68d69b3
- Add __all__ · 21695174
  Dominik Vilsmeier authored Feb 03, 2022
  
  21695174
- Add doc strings to Backend base class · 08887c4d
  Dominik Vilsmeier authored Feb 03, 2022
  
  08887c4d
- Fix not updating tensor data correctly during test · be7d5fe6
  Dominik Vilsmeier authored Feb 03, 2022
  
  be7d5fe6
- Add 'update_tensor_data' function to backends for inplace tensor updates · a76018a1
  Dominik Vilsmeier authored Feb 03, 2022
  
  a76018a1
- Add 'ignore_gradient' context manager to backend classes · 224533ea
  Dominik Vilsmeier authored Feb 03, 2022
  
  224533ea
- Fix unittests not properly discriminating between Tensor and Parameter types · f2cca326
  Dominik Vilsmeier authored Feb 03, 2022
  
  f2cca326
- Fix Numpy backend not enorcing float64 by default · 31ba35c4
  Dominik Vilsmeier authored Feb 03, 2022
  
  31ba35c4
- Add backend method for converting 0-dim tensors to numbers · e8bc02bb
  Dominik Vilsmeier authored Feb 03, 2022
  
  e8bc02bb
- Fix test case using invalid configuration · 7c7f9278
  Dominik Vilsmeier authored Feb 03, 2022
  
  7c7f9278
- Fix test case using invalid configuration · dd830a6c
  Dominik Vilsmeier authored Feb 03, 2022
  
  dd830a6c
- Fix typo · bdfa42be
  Dominik Vilsmeier authored Feb 03, 2022
  
  bdfa42be
- Fix 0-dim arrays being changed to np.float64 when using Numpy backend · 79997bda
  Dominik Vilsmeier authored Feb 03, 2022
  
  79997bda
Feb 02, 2022
- Fix min. version requirement for Numpy · 69b38dd2
  Dominik Vilsmeier authored Feb 02, 2022
  
  69b38dd2
- Fix test case producing NaNs · 00bb733a
  Dominik Vilsmeier authored Feb 02, 2022
  
  00bb733a
- Fix incompatibilities with Numpy backend · 0b80fb8c
  Dominik Vilsmeier authored Feb 02, 2022
  
  0b80fb8c
- Make ModuleProxy compatible with torch.nn.Module · b95f3301
  Dominik Vilsmeier authored Feb 02, 2022
  
  b95f3301
- Allow selection of backend via environment variable · a4adfd1f
  Dominik Vilsmeier authored Feb 02, 2022
  
  a4adfd1f
- Use backend when converting to numpy arrays · 78615087
  Dominik Vilsmeier authored Feb 02, 2022
  
  78615087
- Use backend functions · 884e3622
  Dominik Vilsmeier authored Feb 02, 2022
  
  884e3622
- First check for successful execution, then check specifics · e3fc8e16
  Dominik Vilsmeier authored Feb 02, 2022
  
  e3fc8e16
- Fix AttributeError · a42f3f9d
  Dominik Vilsmeier authored Feb 02, 2022
  
  a42f3f9d
- Use backend · 345f72ba
  Dominik Vilsmeier authored Feb 02, 2022
  
  345f72ba
- Add guards to PyTorch adapters · 76d94fe1
  Dominik Vilsmeier authored Feb 02, 2022
  
  76d94fe1
- Update unittests to use backend · 35da3e9f
  Dominik Vilsmeier authored Feb 02, 2022
  
  35da3e9f
- Update compute module and unittests to use backend · f3d25111
  Dominik Vilsmeier authored Feb 02, 2022
  
  f3d25111
- Add functions to backends · 02e68f21
  Dominik Vilsmeier authored Feb 02, 2022
  
  02e68f21
- Fix annotation with ParameterType · 81ede643
  Dominik Vilsmeier authored Feb 02, 2022
  
  81ede643
- Fix invalid import · ce75a77f
  Dominik Vilsmeier authored Feb 02, 2022
  
  ce75a77f
- Use linalg module from backends · 68873f88
  Dominik Vilsmeier authored Feb 02, 2022
  
  68873f88
- Update minimum version requirement for numpy and torch projects · a487f9ee
  Dominik Vilsmeier authored Feb 02, 2022
  
  a487f9ee
- Add linalg module to backends · a19d2dab
  Dominik Vilsmeier authored Feb 02, 2022
  
  a19d2dab
- There is no need to expose the type variables outside the backends module · e41d30a1
  Dominik Vilsmeier authored Feb 02, 2022
  
  e41d30a1
- Fix type hints incorrectly using the TensorType type variable · 879d7b22
  Dominik Vilsmeier authored Feb 02, 2022
  
  879d7b22
- Fix incorrect usage of Parameter type · 6f8b6f13
  Dominik Vilsmeier authored Feb 02, 2022
  
  6f8b6f13
- Adjust function call to match backend signature · 272cb7a2
  Dominik Vilsmeier authored Feb 02, 2022
  
  272cb7a2