PockEngine: Sparse and Efficient Fine-tuning in a Pocket

Explore a conference talk from MICRO 2023 presenting "PockEngine: Sparse and Efficient Fine-tuning in a Pocket." Delivered by researchers from MIT HAN Lab, this 22-minute presentation introduces an innovative approach to sparse and efficient fine-tuning of machine learning models. Dive into the work of Ligeng Zhu, Lanxiang Hu, Ji Lin, Wei-Chen Wang, Wei-Ming Chen, Chuang Gan, and Song Han as they discuss their groundbreaking research. Learn about the potential applications and implications of PockEngine for improving the efficiency of machine learning model fine-tuning. For those seeking more in-depth information, visit the official PockEngine website at https://pockengine.mit.edu/.

MICRO'23 PockEngine: Sparse and Efficient Fine-tuning in a Pocket