Lightweight Model Pre-Training via Language Guided Knowledge Distillation | IEEE Journals & Magazine | IEEE Xplore