Skip to content
Snippets Groups Projects
Select Git revision
  • master default protected
  • 0.0.5
  • 0.0.4
  • 0.0.3
  • 0.0.2
  • 0.0.1
6 results
You can move around the graph by using the arrow keys.
Created with Raphaël 2.2.03Aug10Apr9516Mar14131227Feb2624Auto-formatmastermasterAdd test_register_hookSave state also for backward passApply crazy state hackDivide through number of projections on backprojectionAssert float().cuda().contiguous() in both forward/backward projectionRemove pyconrad in non-pyconrad testFix backward projectionAdd setter for projection_matricesFix typo in README.rstSkip pyconrad test not only on CI but always if pyconrad is not installedFix indentation of global declarations of PYRO-NN functionsAvoid GNU extension __restrict__ (prevents building with MSVC)Remove some unnecessary statements in setup.pyAvoid copying of projection_matrices on each kernel callTravis Windows: Remove non-working installation of CUDAAvoid copying of projection matrices on each kernel callTravis Windows: random changeTravis: Experiment windowsTry to install cuda on windowsTravis: Try install pip on windowsTravis: add condition to out-commented PyPI deployBuild CUDAExtension with '-03'. Maybe this helps?Travis: use better filenames for wheelsTry to fix CUDA releaseFix generate_shared_object if pystencils_autodiff is not installedTravis: Deactivate not working pypi push for nowAvoid absolute paths in setup.pyRemove --std=c++-14 flag in setup.pyTravis: use underscore after cuda version [ci-skip]Allow failure in TF wrapping for now (is fixed on work PC)Add CUDA 10.2Try to install CUDA version corresponding to CUDA_VERSIONTravis: put code into file that's actually usedTravis: Fix CUDA installationTravis: set CUDA_HOMETravis: Only install `rename` on LinuxFixup: travisTravis: Add /usr/local/cuda/bin to PATHTry to append CUDA_VERSION to artifacts name