M
Marinus van Baalen
Researcher at Qualcomm
Publications - 2
Citations - 176
Marinus van Baalen is an academic researcher from Qualcomm. The author has contributed to research in topics: Rounding & Quadratic unconstrained binary optimization. The author has an hindex of 2, co-authored 2 publications receiving 50 citations.
Papers
More filters
Posted Content
Up or Down? Adaptive Rounding for Post-Training Quantization
TL;DR: AdaRound is proposed, a better weight-rounding mechanism for post-training quantization that adapts to the data and the task loss that outperforms rounding-to-nearest by a significant margin and establishes a new state-of-the-art forPost- training quantization on several networks and tasks.
Proceedings Article
Up or Down? Adaptive Rounding for Post-Training Quantization
TL;DR: AdaRound as discussed by the authors approximates the task loss with a Taylor series expansion and poses the rounding task as a quadratic unconstrained binary optimization problem, which is then simplified to a layer-wise local loss.