rascaline
rascaline copied to clipboard
Computing representations for atomistic machine learning
Thanks to @PicoCentauri and @frostedoyster for finding these bugs! ---- 📚 Documentation preview 📚: https://rascaline--304.org.readthedocs.build/en/304/
We need to add a documentation page explaining what keys, sample/components/properties and gradient components will be produced by the various rascaline calculators. I'm not sure if that should be done...
Hello, I understand a lot of hard work has gone into coding up the CG utils for rascaline. Having this separate util class and functions allows us to do cool...
With the implementation of #298 the spliner classes can now handle all combinations of `radial basis` functions and `atomic densities` we offer 🎉. And, there is already some technical documentation...
the changes I talked about, split into three commits that can be reviewed separately There's just one point where I know feedback is needed, it's for the new API for...
There have been a bunch of build system and CI improvements in metatensor that we should port to rascaline as well - [ ] https://github.com/lab-cosmo/metatensor/pull/382 - [ ] https://github.com/lab-cosmo/metatensor/pull/491 -...
There is an old function, i.e. `rascaline.generate_splines`, which is still mentioned in the documentation. If I understand correctly, it was replaced by `rascaline.utils.RadialIntegralFromFunction`
Hello, as discussed on Slack, here's a feature request: It'd be nice to be able to evaluate a (splined) radial basis at arbitrary points for testing purposes. This would allow...
Here we outline the steps needed to get to a rascaline release. - [ ] Build steps, outlined in #281 - [x] Renaming metadata to match new conventions in metatensor...
## Idea: the format of the output of rascaline.torch.LodeSphericalExpansion() is aligned with e3nn: I am interested in combining rascaline.torch.LodeSphericalExpansion() with e3nn(https://github.com/e3nn/e3nn). Both methods are based on spherical harmonics and satisfy...