Automatic generation of multi-precision multi-arithmetic CNN accelerators for FPGAs
| Autor: | Robert Mullins, Cheng-Zhong Xu, Junyi Liu, Yiren Zhao, Xuan Guo, George A. Constantinides, Erwei Wang, Peter Y. K. Cheung, Xitong Gao |
|---|---|
| Rok vydání: | 2019 |
| Předmět: |
Signal Processing (eess.SP)
FOS: Computer and information sciences Computer Science - Machine Learning Computer science 010501 environmental sciences 01 natural sciences Convolutional neural network Convolution Machine Learning (cs.LG) Software CNN hardware accelerator 0103 physical sciences Hardware Architecture (cs.AR) FOS: Electrical engineering electronic engineering information engineering Arithmetic Latency (engineering) Electrical Engineering and Systems Science - Signal Processing Field-programmable gate array Computer Science - Hardware Architecture Throughput (business) 0105 earth and related environmental sciences 010302 applied physics business.industry Process (computing) Frame rate Auto-generation business |
| Zdroj: | FPT |
| DOI: | 10.17863/cam.46508 |
| Popis: | Modern deep Convolutional Neural Networks (CNNs) are computationally demanding, yet real applications often require high throughput and low latency. To help tackle these problems, we propose Tomato, a framework designed to automate the process of generating efficient CNN accelerators. The generated design is pipelined and each convolution layer uses different arithmetics at various precisions. Using Tomato, we showcase state-of-the-art multi-precision multi-arithmetic networks, including MobileNet-V1, running on FPGAs. To our knowledge, this is the first multi-precision multi-arithmetic auto-generation framework for CNNs. In software, Tomato fine-tunes pretrained networks to use a mixture of short powers-of-2 and fixed-point weights with a minimal loss in classification accuracy. The fine-tuned parameters are combined with the templated hardware designs to automatically produce efficient inference circuits in FPGAs. We demonstrate how our approach significantly reduces model sizes and computation complexities, and permits us to pack a complete ImageNet network onto a single FPGA without accessing off-chip memories for the first time. Furthermore, we show how Tomato produces implementations of networks with various sizes running on single or multiple FPGAs. To the best of our knowledge, our automatically generated accelerators outperform closest FPGA-based competitors by at least 2-4x for lantency and throughput; the generated accelerator runs ImageNet classification at a rate of more than 3000 frames per second. To be published in International Conference on Field Programmable Technology 2019 |
| Databáze: | OpenAIRE |
| Externí odkaz: |
|