TY  - RPRT
TI  - Vision Models Can Be Efficiently Specialized via Few-Shot Task-Aware Compression
AU  - Denis Kuznedelev
AU  - Soroush Tabesh
AU  - Kimia Noorbakhsh
AU  - Elias Frantar
AU  - Sara Beery
AU  - Eldar Kurtic
AU  - Dan Alistarh
PY  - 2023
UR  - https://arxiv.org/abs/2303.14409
ID  - 2303.14409
ER  -