Abstract | ||
---|---|---|
It is always well believed that Binary Neural Networks (BNNs) could drastically accelerate the inference efficiency by replacing the arithmetic operations in float-valued Deep Neural Networks (DNNs) with bit-wise operations. Nevertheless, there has not been open-source implementation in support of this idea on low-end ARM devices (e.g., mobile phones and embedded devices). In this work, we propose daBNN --- a super fast inference framework that implements BNNs on ARM devices. Several speed-up and memory refinement strategies for bit-packing, binarized convolution, and memory layout are uniquely devised to enhance inference efficiency. Compared to the recent open-source BNN inference framework, BMXNet, our daBNN is 7x~23x faster on a single binary convolution, and about 6x faster on Bi-Real Net 18 (a BNN variant of ResNet-18). The daBNN is a BSD-licensed inference framework, and its source code, sample projects and pre-trained models are available on-line: https://github.com/JDAI-CV/dabnn.
|
Year | DOI | Venue |
---|---|---|
2019 | 10.1145/3343031.3350534 | Proceedings of the 27th ACM International Conference on Multimedia |
Keywords | Field | DocType |
binary neural networks, machine learning, open source | Computer vision,Computer science,Inference,Binary neural network,Artificial intelligence | Conference |
ISBN | Citations | PageRank |
978-1-4503-6889-6 | 7 | 0.44 |
References | Authors | |
0 | 5 |
Name | Order | Citations | PageRank |
---|---|---|---|
Jianhao Zhang | 1 | 7 | 1.12 |
Yingwei Pan | 2 | 357 | 23.66 |
Ting Yao | 3 | 842 | 52.62 |
He Zhao | 4 | 11 | 1.14 |
Tao Mei | 5 | 4702 | 288.54 |