Sökning: onr:"swepub:oai:DiVA.org:miun-43421" >
Neural Network Comp...
Neural Network Compression Through Shunt Connections and Knowledge Distillation for Semantic Segmentation Problems
-
Haas, B. (författare)
-
Wendt, A. (författare)
-
Jantsch, A. (författare)
-
visa fler...
-
Wess, M. (författare)
-
visa färre...
- 2021-06-22
- 2021
- Engelska.
-
Ingår i: IFIP Advances in Information and Communication Technology. - Cham : Springer Science and Business Media Deutschland GmbH. - 9783030791490 ; , s. 349-361
- Relaterad länk:
-
https://urn.kb.se/re...
-
visa fler...
-
https://doi.org/10.1...
-
visa färre...
Abstract
Ämnesord
Stäng
- Employing convolutional neural network models for large scale datasets represents a big challenge. Especially embedded devices with limited resources cannot run most state-of-the-art model architectures in real-time, necessary for many applications. This paper proves the applicability of shunt connections on large scale datasets and narrows this computational gap. Shunt connections is a proposed method for MobileNet compression. We are the first to provide results of shunt connections for the MobileNetV3 model and for segmentation tasks on the Cityscapes dataset, using the DeeplabV3 architecture, on which we achieve compression by 28%, while observing a 3.52 drop in mIoU. The training of shunt-inserted models are optimized through knowledge distillation. The full code used for this work will be available online. © 2021, IFIP International Federation for Information Processing.
Nyckelord
- Accuracy
- CIFAR
- Cityscapes
- DeepLab
- Embedded machine learning
- Knowledge distillation
- Latency
- Machine learning
- MobileNet
- Optimization
- Shunt connections
- Convolutional neural networks
- Distillation
- Large dataset
- Network architecture
- Semantic Web
- Semantics
- Embedded device
- Large-scale datasets
- Network compression
- Real time
- Semantic segmentation
- State of the art
- Distilleries
Publikations- och innehållstyp
- ref (ämneskategori)
- kon (ämneskategori)
Hitta via bibliotek
Till lärosätets databas