parallelizable-linear-rnns

Linear RNN architecture enabling efficient parallel computation over directed acyclic graphs with stabilization modes for long-range dependencies without sequential traversal.

Independent GitHub snapshot · Automated classification · Original documentation governs