A Controlled Evaluation with a Last-Stage Multiscale Variant
Journal of Big Data and Computing · 2026-09 · 4 (3) · 107-112
DOI: 10.62517/jbdc.202601314 ↗Kurz gefasst
Fachartikel in der Zeitschrift Journal of Big Data and Computing, 2026-09. Verfasst von Zhaoyang Chen und 1 weiteren Autor:innen. Die Arbeit nennt 14 Literaturangaben.
Abstract (Originalsprache)
Efficiency claims for compact convolutional networks are commonly supported by parameter counts and floating-point operation estimates, although neither metric directly measures execution time. This study examines how accuracy, static complexity, and observed GPU runtime align for three established lightweight networks and one controlled architectural variant. MobileNetV2, MobileNetV3-Small, ShuffleNetV2, and a MobileNetV3 variant with a parallel multi-scale module in its final inverted-residual block were trained from scratch on CIFAR-10 and CIFAR-100. A fixed data split and training recipe were used throughout, with three independent random seeds per model and dataset, yielding 24 formal runs. ShuffleNetV2 produced the highest mean test accuracy on both datasets, reaching 93.76% on CIFAR-10 and 74.24% on CIFAR-100. MobileNetV2 was fastest in the hardware benchmark despite requiring approximately four times the FLOPs of MobileNetV3. The multi-scale intervention increased CIFAR-10 accuracy by only 0.39 percentage points while adding 45.09% more parameters and 43.89% more FLOPs. On CIFAR-100, it reduced accuracy by 4.18 percentage points and also lowered throughput. The findings show that low FLOPs should not be treated as a sufficient proxy for runtime and that a multi-scale operator can become counterproductive when its placement and fusion cost disturb an already compact representation.
Zitieren
APA
Chen, Z., & Li, T. (2026). A Controlled Evaluation with a Last-Stage Multiscale Variant. Journal of Big Data and Computing, 4(3), 107-112. https://doi.org/10.62517/jbdc.202601314
Vancouver
Chen Z, Li T. A Controlled Evaluation with a Last-Stage Multiscale Variant. Journal of Big Data and Computing. 2026;4(3):107-112. doi:10.62517/jbdc.202601314
BibTeX
@article{Chen2026,
author = {Chen, Zhaoyang and Li, Tuchao},
title = {A Controlled Evaluation with a Last-Stage Multiscale Variant},
journal = {Journal of Big Data and Computing},
year = {2026},
volume = {4},
number = {3},
pages = {107--112},
doi = {10.62517/jbdc.202601314}
}
EUVIMED