| Presentation | 2022-06-08 A Compact High-Speed CNN Implementation based on Redundant Computational Analysis and FPGA Acceleration Li Qi, Li Hengyi, Meng Lin, |
|---|---|
| PDF Download Page |  PDF download Page Link |
| Abstract(in Japanese) | (See Japanese page) |
| Abstract(in English) | Convolutional Neural Networks (CNNs) have achieved high performance and are widely used in various applications. However, CNN's are computational-intensive and resource-consuming, causing the development of CNN applications is limited especially in the embedded systems. Therefore, we propose a dynamic CNN pruning method based on redundant computational analysis. The proposal aims to realize model compression within the setting performance degradation through dynamic iterative channel pruning. Experimental results show the proposal reduces about 80% of parameters and FLOPS. Furthermore, the compacted CNN model is implemented on the FPGA and achieved about 40% speedup in inference time. |
| Keyword(in Japanese) | (See Japanese page) |
| Keyword(in English) | Convolutional neural network / Channel pruning / Redundant calculation analysis / FPGA |
| Paper # | RECONF2022-21 |
| Date of Issue | 2022-05-31 (RECONF) |
| Conference Information | |
| Committee | RECONF |
|---|---|
| Conference Date | 2022/6/7(2days) |
| Place (in Japanese) | (See Japanese page) |
| Place (in English) | CCS, Univ. of Tsukuba |
| Topics (in Japanese) | (See Japanese page) |
| Topics (in English) | Reconfigurable system, etc. |
| Chair | Kentaro Sano(RIKEN) |
| Vice Chair | Yoshiki Yamaguchi(Tsukuba Univ.) / Tomonori Izumi(Ritsumeikan Univ.) |
| Secretary | Yoshiki Yamaguchi(NEC) / Tomonori Izumi(Toyohashi Univ. of Tech.) |
| Assistant | Yukitaka Takemura(INTEL) / Yasunori Osana(Ryukyu Univ.) |
| Paper Information | |
| Registration To | Technical Committee on Reconfigurable Systems |
|---|---|
| Language | JPN |
| Title (in Japanese) | (See Japanese page) |
| Sub Title (in Japanese) | (See Japanese page) |
| Title (in English) | A Compact High-Speed CNN Implementation based on Redundant Computational Analysis and FPGA Acceleration |
| Sub Title (in English) | |
| Keyword(1) | Convolutional neural network |
| Keyword(2) | Channel pruning |
| Keyword(3) | Redundant calculation analysis |
| Keyword(4) | FPGA |
| 1st Author's Name | Li Qi |
| 1st Author's Affiliation | Ritsumeikan University(Ritsumeikan Univ.) |
| 2nd Author's Name | Li Hengyi |
| 2nd Author's Affiliation | Ritsumeikan University(Ritsumeikan Univ.) |
| 3rd Author's Name | Meng Lin |
| 3rd Author's Affiliation | Ritsumeikan University(Ritsumeikan Univ.) |
| Date | 2022-06-08 |
| Paper # | RECONF2022-21 |
| Volume (vol) | vol.122 |
| Number (no) | RECONF-60 |
| Page | pp.pp.89-94(RECONF), |
| #Pages | 6 |
| Date of Issue | 2022-05-31 (RECONF) |