Abstract | ||
---|---|---|
Recently, convolutional networks have achieved great successes in the field of computer vision. In order to improve the efficiency of convolutional networks, large amount of solutions focusing on training algorithms and parallelism strategies have been proposed. In this paper, a novel algorithm based on look-up table is proposed to speed up convolutional networks with small filters by applying GPU. By transforming multiplication operations in the convolution computation to some table-based summation operations, the overhead of convolution computation can be reduced largely. The process of creating table and looking up table is very appropriate for parallelization on GPU. Experiment results show that the proposed approaches can improve the speed of convolution computation by 20%-30%, compared with state-of-the-art existing works. |
Year | DOI | Venue |
---|---|---|
2015 | 10.1109/HPCC-CSS-ICESS.2015.218 | HPCC/CSS/ICESS |
Keywords | Field | DocType |
Deep Learning, Convolutional Network, Parallel Computation, GPU | Kernel (linear algebra),Digital filter,Convolution,Computer science,Parallel computing,Multiplication,Artificial intelligence,Deep learning,Computation,Speedup | Conference |
ISSN | Citations | PageRank |
2576-3504 | 0 | 0.34 |
References | Authors | |
9 | 5 |
Name | Order | Citations | PageRank |
---|---|---|---|
Wenbin Jiang | 1 | 7 | 2.90 |
Yiming Chen | 2 | 51 | 9.12 |
Hai Jin | 3 | 6544 | 644.63 |
Bin Luo | 4 | 802 | 107.57 |
Ye Chi | 5 | 6 | 1.83 |