This is the PyTorch implementation of SmRNet: Scalable Multiresolution Feature Extraction Network. Serving as a versatile backbone, the network integrates the discrete wavelet transform (DWT) and its inverse (IDWT) to cater to various computer vision tasks, including detection, classification, and tracking.
If you find this work useful, please cite:
@INPROCEEDINGS{10389571,
author={Alaba, Simegnew Yihunie and Ball, John E.},
booktitle={2023 International Conference on Electrical, Computer and Energy Technologies (ICECET)},
title={SmRNet: Scalable Multiresolution Feature Extraction Network},
year={2023},
volume={},
number={},
pages={1-6},
doi={10.1109/ICECET58911.2023.10389571}}
git clone https://github.com/Simeon340703/SmRNet.git
Install PyTorch and related.
#The default batch size is 128. model choices=['SmRNet_l', 'SmRNet_m', 'SmRNet_s']. dataset choices=['cifar10', 'cifar100'],
python main.py --batch-size 128 --lr 0.1 --model SmRNet_s --dataset cifar100 --epochs 100
- Add Object Detection
- Add Semantic Segmentation