On the impact of the parametrization of deep convolutional neural networks on post-training quantization
Fuente:
arXiv
Salvato in:
| Autori principali: | , , |
|---|---|
| Natura: | Preprint |
| Pubblicazione: |
2025
|
| Soggetti: | |
| Accesso online: | |
| Tags: |
Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
|
| _version_ | 1866912639973588992 |
|---|---|
| author | Houache, Samy Aujol, Jean François Traonmilin, Yann |
| author_facet | Houache, Samy Aujol, Jean François Traonmilin, Yann |
| contents | 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. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2502_01156 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | On the impact of the parametrization of deep convolutional neural networks on post-training quantization Houache, Samy Aujol, Jean François Traonmilin, Yann Information Theory 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. |
| title | On the impact of the parametrization of deep convolutional neural networks on post-training quantization |
| topic | Information Theory |
| url | https://arxiv.org/abs/2502.01156 |