pytorch

e8f916bf - Implement adding bias vector into structured sparse linear operator