[Debichem-devel] Bug#790803: ITP: neural -- machine-learning for atomistics

Andreas Tille andreas at an3as.eu
Thu Jul 16 11:53:09 UTC 2015


Hi Graham,

I think this ITP is relevant for Debian Science and DebiChem.  I guess
you intend to maintain it in one of these teams.  Please make sure it
will be added to the according Blends tasks (I'm fine if you tell me to
what task what binary package should be added).

It would be nice if you would CC the relevant teams in the beginning
(and sorry if I missed the announcement).

Thanks for this ITP

     Andreas.

On Wed, Jul 01, 2015 at 10:51:15PM +0200, Graham Inggs wrote:
> Package: wnpp
> Severity: wishlist
> Owner: Graham Inggs <graham at nerve.org.za>
> X-Debbugs-CC: debian-devel at lists.debian.org
> 
> * Package name    : neural
>   Version         : 1.0
>   Upstream Author : Andrew Peterson, Alireza Khorshidi
> * URL             : https://bitbucket.org/andrewpeterson/neural
> * License         : GPL-3.0+
>   Programming Lang: Python
>   Description     : Machine Learning for Atomistics
>  Neural is an open-source code designed to easily bring machine-learning to
>  atomistic calculations. This allows one to predict (or really, interpolate)
>  calculations on the potential energy surface, by optimizing a neural network
>  representation of a "training set" of atomic images. The code works by
>  learning from any other calculator (usually DFT) that can provide energy as
>  a function of atomic coordinates. In theory, these predictions can take place
>  with arbitrary accuracy approaching that of the original calculator.
>  .
>  Neural is designed to integrate closely with the Atomic Simulation
>  Environment (ASE). As such, the interface is in pure python, although several
>  compute-heavy parts of the underlying code also have fortran versions to
>  accelerate the calculations. The close integration with ASE means that any
>  calculator that works with ASE ─ including EMT, GPAW, DACAPO, VASP, NWChem,
>  and Gaussian ─ can easily be used as the parent method.
> 
> 
> --
> To UNSUBSCRIBE, email to debian-devel-REQUEST at lists.debian.org
> with a subject of "unsubscribe". Trouble? Contact listmaster at lists.debian.org
> Archive: https://lists.debian.org/CAM8zJQs+C4KMO1YbOp3SnfoqP3NV6wX8wkK4Zo5YSyA+J76hug@mail.gmail.com
> 
> 

-- 
http://fam-tille.de



More information about the Debichem-devel mailing list