AMC: AutoML for Model Compression and Acceleration on Mobile Devices
Yihui He, Ji Lin, Zhijian Liu, Hanrui Wang, Li-Jia Li, Song Han
Code Available — Be the first to reproduce this paper.
ReproduceCode
- github.com/mit-han-lab/amcOfficialpytorch★ 0
- github.com/mit-han-lab/once-for-allpytorch★ 1,944
- github.com/ito-rafael/once-for-all-2pytorch★ 3
- github.com/mit-han-lab/haqpytorch★ 0
- github.com/seulkiyeom/once-for-allpytorch★ 0
- github.com/mit-han-lab/ProxylessNASpytorch★ 0
- github.com/NervanaSystems/distillerpytorch★ 0
- github.com/songhan/SqueezeNet-Residualnone★ 0
- github.com/songhan/DSDpytorch★ 0
- github.com/AhmadQasim/proxylessnas-densepytorch★ 0
Abstract
Model compression is a critical technique to efficiently deploy neural network models on mobile devices which have limited computation resources and tight power budgets. Conventional model compression techniques rely on hand-crafted heuristics and rule-based policies that require domain experts to explore the large design space trading off among model size, speed, and accuracy, which is usually sub-optimal and time-consuming. In this paper, we propose AutoML for Model Compression (AMC) which leverage reinforcement learning to provide the model compression policy. This learning-based compression policy outperforms conventional rule-based compression policy by having higher compression ratio, better preserving the accuracy and freeing human labor. Under 4x FLOPs reduction, we achieved 2.7% better accuracy than the handcrafted model compression policy for VGG-16 on ImageNet. We applied this automated, push-the-button compression pipeline to MobileNet and achieved 1.81x speedup of measured inference latency on an Android phone and 1.43x speedup on the Titan XP GPU, with only 0.1% loss of ImageNet Top-1 accuracy.