librascal
librascal copied to clipboard
lab-cosmo
A scalable and versatile library to generate representations for atomic-scale learning
In the tests the cutoff of neighbourlist manager and of the representation calculator are mismatched, this does not cause any error but is invalid behavior which can cause at some...
This is a bit of a nag really, but const correctness has been an issue ever since the start; it's never been easy with our complex network of heavy custom...
We should enforce PEP8 conformity using flake8 and run it in the CI. This corresponds to update the codebase making sure it is PEP8 compatible, add a flake8 target in...
As discussed with @agoscinski and @Luthaf, we need to look at the compiler flags, especially e.g. an updated version of the `-Weffc++`. Opened an issue so we can discuss it...
I am not questioning the numbers or intend. What I am trying to say: can we refactor this piece of code in terms of the individual contributions (guess that is...
Found in PR #256 : some of the python files are not autopepped when running `make pretty-python`. Possible reasons: - files not registered - autopep not configured correctly -?
Writing an issue for myself as a reminder to check these things: For gradients: - store intermediate results without summation (needed for kernel partial derivatives) Behler Angular symmetry functions, implementation...
Compiling with gcc-9 with -fsanitize=undefined compiler flag (see branch `bugfix/undefined` leads to several reference binding to null pointer. I fixed the ones related to spherical_expansion in the `feat/sparse_kernel` branch so...
Following the discussion on #254 , I am posting an issue here for a possible future improvement: The `calculator_spherical_*` does some sorting/adapting on its own during computation. And I think...
This is a draft PR to provide a pipeline for adding the tutorials to master. The rebase became unwieldy, so @Luthaf suggested a new branch from master with only the...
