TP7a.4

Parameter Efficient Fine-tuning of Transformer-based Language Models Using Dataset Pruning

Sayed Mohammadreza Tayaranian Hosseini, Seyyed Hasan Mozafari, James Clark, Brett Meyer, Warren Gross, McGill University, Canada
Session:
TP7a: Hardware and Software Approaches for Efficient Deep Neural Networks Lecture
Track:
Architectures and Implementation
Location:
Oak Shelter
Presentation Time:
Tue, 29 Oct, 14:45 - 15:10 PT (UTC -7)
Presentation
Discussion
Resources
No resources available.
Session TP7a
TP7a.1: Investigation and Rethinking of Dynamic Quantization for Deep Neural Networks
Erjing Luo, Jie Han, University of Alberta, Canada
TP7a.2: Every Bit Matters: A Hardware/Software Approach for Enabling More Powerful Machine Learning Models
Enrique Torres Sannchez, Milos Nikolic, Kareem Ibrahim, Alberto Delmas Lascorz, Ali Hadi Zadeh, Jiahui Wang, University of Toronto, Canada; Ameer Abdelhadi, McMaster University, Canada; Mostafa Mahmoud, Christina Giannoula, Andreas Moshovos, University of Toronto, Canada
TP7a.3: SAMSON: Sharpness-Aware Minimization Scaled by Outlier Normalization for Improving Robustness on Noisy DNN Accelerators
Sébastien Henwood, Polytechnique Montreal, Canada; Gonçalo Mordido, Mila - Quebec AI Institute, Canada; Yvon Savaria, Polytechnique Montreal, Canada; Sarath Chandar, Mila - Quebec AI Institute, Canada; François Leduc-Primeau, Polytechnique Montreal, Canada
TP7a.4: Parameter Efficient Fine-tuning of Transformer-based Language Models Using Dataset Pruning
Sayed Mohammadreza Tayaranian Hosseini, Seyyed Hasan Mozafari, James Clark, Brett Meyer, Warren Gross, McGill University, Canada
Contacts