Adaptive quantization with mixed-precision based on low-cost proxy
| Autor: | Chen, Junzhe, Yang, Qiao, Tian, Senmao, Zhang, Shunli |
|---|---|
| Rok vydání: | 2024 |
| Předmět: | |
| Zdroj: | ICASSP2024 |
| Druh dokumentu: | Working Paper |
| DOI: | 10.1109/ICASSP48485.2024.10447866 |
| Popis: | It is critical to deploy complicated neural network models on hardware with limited resources. This paper proposes a novel model quantization method, named the Low-Cost Proxy-Based Adaptive Mixed-Precision Model Quantization (LCPAQ), which contains three key modules. The hardware-aware module is designed by considering the hardware limitations, while an adaptive mixed-precision quantization module is developed to evaluate the quantization sensitivity by using the Hessian matrix and Pareto frontier techniques. Integer linear programming is used to fine-tune the quantization across different layers. Then the low-cost proxy neural architecture search module efficiently explores the ideal quantization hyperparameters. Experiments on the ImageNet demonstrate that the proposed LCPAQ achieves comparable or superior quantization accuracy to existing mixed-precision models. Notably, LCPAQ achieves 1/200 of the search time compared with existing methods, which provides a shortcut in practical quantization use for resource-limited devices. Comment: accepted by icassp2024 |
| Databáze: | arXiv |
| Externí odkaz: |
|