Towards Clip-Free Quantized Super-Resolution Networks: How to Tame Representative Images
Alperen Kalay (Aselsan Research),* Bahri Batuhan Bilecen (Aselsan Research), Mustafa Ayazoglu (Aselsan Research)
The 34th British Machine Vision Conference
Abstract
Super-resolution (SR) networks have been investigated for a while, with their mobile and lightweight versions gaining noticeable popularity recently. Quantization, the procedure of decreasing the precision of network parameters (mostly FP32 to INT8), is also utilized in SR networks for establishing mobile compatibility. This study focuses on a very important but mostly overlooked post-training quantization(PTQ) step: the representative dataset (RD), which adjusts the quantization range for PTQ. We propose a novel pipeline backed up with extensive experimental justifications to cleverly augment RD images by only using outputs of the FP32 model. Using the proposed pipeline for RD, we can successfully eliminate unwanted clipped activation layers, which nearly all mobile SR methods utilize to make the model more robust to PTQ in return for a large overhead in runtime. Removing clipped activations with our method significantly benefits overall increased stability, decreased inference runtime up to 54% on some SR models, better visual quality results compared to INT8 clipped models - and outperforms even some FP32 non-quantized models as well, both in runtime and visual quality.Video
Citation
@inproceedings{Kalay_2023_BMVC,
author = {Alperen Kalay and Bahri Batuhan Bilecen and Mustafa Ayazoglu},
title = {Towards Clip-Free Quantized Super-Resolution Networks: How to Tame Representative Images},
booktitle = {34th British Machine Vision Conference 2023, {BMVC} 2023, Aberdeen, UK, November 20-24, 2023},
publisher = {BMVA},
year = {2023},
url = {https://papers.bmvc2023.org/0775.pdf}
}
Copyright © 2023
The British Machine Vision Conference is organised by