Data Quality-aware Mixed-precision Quantization via Hybrid Reinforcement Learning
Wang, Yingchun, Guo, Jingcai, Guo, Song, Zhang, Weizhan
–arXiv.org Artificial Intelligence
Mixed-precision quantization mostly predetermines the model bit-width settings before actual training due to the non-differential bit-width sampling process, obtaining sub-optimal performance. Worse still, the conventional static quality-consistent training setting, i.e., all data is assumed to be of the same quality across training and inference, overlooks data quality changes in real-world applications which may lead to poor robustness of the quantized models. In this paper, we propose a novel Data Quality-aware Mixed-precision Quantization framework, dubbed DQMQ, to dynamically adapt quantization bit-widths to different data qualities. The adaption is based on a bit-width decision policy that can be learned jointly with the quantization training. Concretely, DQMQ is modeled as a hybrid reinforcement learning (RL) task that combines model-based policy optimization with supervised quantization training. By relaxing the discrete bit-width sampling to a continuous probability distribution that is encoded with few learnable parameters, DQMQ is differentiable and can be directly optimized end-to-end with a hybrid optimization target considering both task performance and quantization benefits. Trained on mixed-quality image datasets, DQMQ can implicitly select the most proper bit-width for each layer when facing uneven input qualities. Extensive experiments on various benchmark datasets and networks demonstrate the superiority of DQMQ against existing fixed/mixed-precision quantization methods.
arXiv.org Artificial Intelligence
Feb-9-2023
- Country:
- North America
- United States
- Washington > King County
- Seattle (0.04)
- Utah > Salt Lake County
- Salt Lake City (0.04)
- Nevada > Clark County
- Las Vegas (0.04)
- Louisiana > Orleans Parish
- New Orleans (0.04)
- California > Los Angeles County
- Long Beach (0.04)
- Washington > King County
- Puerto Rico > San Juan
- San Juan (0.04)
- Canada > British Columbia
- United States
- Europe
- France (0.04)
- Germany > Bavaria
- Upper Bavaria > Munich (0.04)
- Asia
- Macao (0.04)
- South Korea > Seoul
- Seoul (0.04)
- China
- Shaanxi Province > Xi'an (0.04)
- Hong Kong (0.04)
- Africa > Ethiopia
- Addis Ababa > Addis Ababa (0.04)
- North America
- Genre:
- Research Report (0.82)
- Technology: