Back to Search
Start Over
Pruning In Time (PIT): A Lightweight Network Architecture Optimizer for Temporal Convolutional Networks
- Source :
- DAC
- Publication Year :
- 2021
- Publisher :
- IEEE, 2021.
-
Abstract
- Temporal Convolutional Networks (TCNs) are promising Deep Learning models for time-series processing tasks. One key feature of TCNs is time-dilated convolution, whose optimization requires extensive experimentation. We propose an automatic dilation optimizer, which tackles the problem as a weight pruning on the time-axis, and learns dilation factors together with weights, in a single training. Our method reduces the model size and inference latency on a real SoC hardware target by up to 7.4x and 3x, respectively with no accuracy drop compared to a network without dilation. It also yields a rich set of Pareto-optimal TCNs starting from a single model, outperforming hand-designed solutions in both size and accuracy.
- Subjects :
- FOS: Computer and information sciences
Computer Science - Machine Learning
Network architecture
Artificial neural network
Computer science
business.industry
Deep learning
Neural Architecture Search
Deep Learning
Edge Computing
Temporal Convolutional Networks
Machine Learning (cs.LG)
Convolution
Set (abstract data type)
Dilation (metric space)
Feature (computer vision)
Pruning (decision trees)
Artificial intelligence
business
Algorithm
Subjects
Details
- Database :
- OpenAIRE
- Journal :
- 2021 58th ACM/IEEE Design Automation Conference (DAC)
- Accession number :
- edsair.doi.dedup.....b2274b8db5d89e2d84554283110d962f