Accelerating attention through gradient-based learned runtime pruning.

Zheng Li, Soroush Ghodrati, Amir Yazdanbakhsh, Hadi Esmaeilzadeh, Mingu Kang

Browse the full ISCA paper archive.