Evaluating and optimizing neural network models with neuromorphic capable and non-capable hardware
The main objective of this project is to evaluate and optimize Spiking Neural Network with the Novena Chip to achieve high accuracy, low processing time, and low power consumption. The integrated pair (Spiking Neural Network with Novena) will be benchmarked against other conventional convolution neu...
Saved in:
| Main Author: | |
|---|---|
| Other Authors: | |
| Format: | Final Year Project |
| Language: | English |
| Published: |
Nanyang Technological University
2021
|
| Subjects: | |
| Online Access: | https://hdl.handle.net/10356/149657 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Institution: | Nanyang Technological University |
| Language: | English |
| Summary: | The main objective of this project is to evaluate and optimize Spiking Neural Network with the Novena Chip to achieve high accuracy, low processing time, and low power consumption. The integrated pair (Spiking Neural Network with Novena) will be benchmarked against other conventional convolution neural networks running on non-neuromorphic hardware. The conventional convolution neural networks used in this paper will be ResNet-50, Inception V4, and MobileNet. The non-neuromorphic hardware used will be Nvidia’s NanoJetson, Raspberry Pi 4B with Intel’s Neural Compute Stick 2, Raspberry Pi 4B with Coral’s USB Accelerator, and ASUS Tinker Edge T. All experiments will be making use of the same dataset for both visual and audio component. |
|---|