Experimental Comparison of Light-Weight and Deep CNN Models Across Diverse Datasets
Md. Hefzul Hossain Papon, Shadman Rabby · 2026
Read Experimental Comparison of Light-Weight and Deep CNN Models Across Diverse Datasets by Md. Hefzul Hossain Papon, Shadman Rabby (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2601-03463.txt.
Description
Md. Hefzul Hossain Papon and Shadman Rabby compare light CNN image models with deep ones across varied datasets, asking how much size really matters. Their tests span smart city surveillance and crop variety classification, two very different uses. The work sets a shared, reproducible benchmark built from several Bangladeshi vision datasets.
Their main finding is that a shallow network, when well regularized, can rival much larger models. Regularization means techniques that stop a model from memorizing its training data. That result matters most where budgets and hardware are thin. A small model that runs on modest machines can serve farms, towns, and schools that cannot rent large graphics processors or buy pretrained systems.
Catalog record
| CREATOR | Md. Hefzul Hossain Papon, Shadman Rabby |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 40 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2601-03463