In principle we need Angles2Coords and other FullAtomModel layers working on GPU. The only complication is that we use recursive depth-first graph traversal which is not suited for vector processors. If this traversal can be effectively unrolled the computation time seems to benefit immensely.
In principle we need Angles2Coords and other FullAtomModel layers working on GPU. The only complication is that we use recursive depth-first graph traversal which is not suited for vector processors. If this traversal can be effectively unrolled the computation time seems to benefit immensely.