On the Impact of the Parametrization of Deep Convolutional Neural Networks on Post-Training Quantization
Abstract
This paper introduces novel theoretical approximation bounds for the output of quantized neural networks, with a focus on convolutional neural networks (CNN). By considering layerwise parametrization and focusing on the quantization of weights, we provide bounds that gain several orders of magnitude compared to state-of-the-art results on classical deep convolutional neural networks such as MobileNetV2 or ResNets. These gains are achieved by improving the behaviour of the approximation bounds with respect to the depth parameter, which has the most impact on the approximation error induced by quantization. To complement our theoretical result, we provide a numerical exploration of our bounds on MobileNetV2 and ResNets.
Cite
Text
Houache et al. "On the Impact of the Parametrization of Deep Convolutional Neural Networks on Post-Training Quantization." Transactions on Machine Learning Research, 2026.Markdown
[Houache et al. "On the Impact of the Parametrization of Deep Convolutional Neural Networks on Post-Training Quantization." Transactions on Machine Learning Research, 2026.](https://mlanthology.org/tmlr/2026/houache2026tmlr-impact/)BibTeX
@article{houache2026tmlr-impact,
title = {{On the Impact of the Parametrization of Deep Convolutional Neural Networks on Post-Training Quantization}},
author = {Houache, Samy and Aujol, Jean-François and Traonmilin, Yann},
journal = {Transactions on Machine Learning Research},
year = {2026},
url = {https://mlanthology.org/tmlr/2026/houache2026tmlr-impact/}
}