Automated Models for Predicting Software Defects in Hybrid MPI and OpenMP Parallel Programs Using Deep Learning