QLTH is a 2-step model compression scheme that applies Post-Training Quantization on the ”winning tickets” or ”matching” derived by iterative magnitude prunning.
Geek Repo:Geek Repo
Github PK Tool:Github PK Tool
