Michele Ceriotti
Michele Ceriotti
Just had a quick browse through the commit diff - seems strange that dftb+ doesn't make the cut - the main publication is from 2007 and is listed on GScholar...
Hello! @nickjbrowning did this PR actually happen? Can we close this issue?
Living on the edge. Will only compile with cmake_nightly.
Looks good - but I get some strange timings on e3nn with small lmax and I wonder if there's something fishy. with ``` python benchmark.py -l 5 -s 100 -t...
I agree. I think that we need to start a discussion with @nickjbrowning in this respect.
I think the main issue will be packaging, but it's worth starting at least a draft PR
If you can / are willing to "harmonize" the API, I think it'd be interesting to have it in the same repo. The idea here is to re-use as much...
Yeah let's hear from @Luthaf and @frostedoyster . I think it'd be really good to consolidate the infrastructure we all use, and this is an easy starting point. When I...
> I guess for large maxL it won't matter anyhow but for small maxL it might have an impact on performance. > > For modelling I would want to provide...
Re license I don't mind dual license. Also better to do it now than after we have a bunch of contributors we can't track. Re the normalization, I suggest that...