Browse files


add abstract, authorship and license
  • Loading branch information...
dalgu90 committed Jun 28, 2017
1 parent b5583b1 commit 4a6917aaefd9da892fdcb9a96404553bbce39664
Showing with 25 additions and 3 deletions.
  1. +25 −3
@@ -4,14 +4,16 @@ TensorFlow implementation of SplitNet: Learning to Semantically Split Deep Netwo
![SplitNet Concept](
- Juyong Kim\*(SNU), Yookoon Park\*(SNU), Gunhee Kim(SNU), and Sung Ju Hwang(UNIST) (*: Equal contributions)
- Juyong Kim\*(SNU), Yookoon Park\*(SNU), Gunhee Kim(SNU), and Sung Ju Hwang(UNIST) (\*: Equal contributions)
We propose a novel deep neural network that is both lightweight and effectively structured for model parallelization. Our network, which we name as *SplitNet*, automatically learns to split the network weights into either a set or a hierarchy of multiple groups that use disjoint sets of features, by learning both the class-to-group and feature-to-group assignment matrices along with the network weights. This produces a tree-structured network that involves no connection between branched subtrees of semantically disparate class groups. SplitNet thus greatly reduces the number of parameters and required computations, and is also embarrassingly model-parallelizable at test time, since the evaluation for each subnetwork is completely independent except for the shared lower layer weights that can be duplicated over multiple processors, or assigned to a separate processor. We validate our method with two different deep network models (ResNet and AlexNet) on two datasets (CIFAR-100 and ILSVRC 2012) for image classification, on which our method obtains networks with significantly reduced number of parameters while achieving comparable or superior accuracies over original full deep networks, and accelerated test speed with multiple GPUs.
## Prerequisite
1. TensorFlow
2. Train/val/test split of CIFAR-100 dataset(please run `python`)
<b>How To Run</b>
## How To Run
# Clone the repo.
@@ -30,3 +32,23 @@ python
# To evaluate
## Acknowledgement
This work was supported by Samsung Research Funding Center of Samsung Electronics under Project Number SRFC-IT150203.
## Authors
[Juyong Kim](<sup>*1</sup>, Yookoon Park<sup>*1</sup>, [Gunhee Kim](<sup>1</sup>, and [Sung Ju Hwang](<sup>2</sup>
1: [Vision and Learning Lab]( @ Computer Science and Engineering, Seoul National University, Seoul, Korea
2: [MLVR Lab]( @ School of Electrical and Computer Engineering, UNIST, Ulsan, South Korea
\*: Equal contribution
## License
MIT license
If you find any problem, please feel free to contact to the authors. :^)

0 comments on commit 4a6917a

Please sign in to comment.