APT: Adaptive Pruning and Tuning Pretrained Language Models for Efficient Training and Inference.

Bowen Zhao, Hannaneh Hajishirzi, Qingqing Cao

Browse the full ICML paper archive.