Lightweight CNNs benchmarked for accuracy and efficiency

A new study published on arXiv provides a reproducible benchmark for lightweight Convolutional Neural Networks (CNNs), comparing seven established architectures across CIFAR-10, CIFAR-100, and Tiny ImageNet datasets. The research evaluated models based on accuracy, parameter count, storage, and computational operations under a unified fine-tuning protocol. EfficientNetV2-S achieved the highest top-1 accuracy, while EfficientNet-B0 offered a strong balance of performance and efficiency, using significantly fewer parameters and operations. The study also highlighted the substantial benefit of ImageNet pretraining, particularly on larger datasets like CIFAR-100 and Tiny ImageNet. AI

IMPACT Provides a clear reference for selecting efficient CNNs, aiding developers in resource-constrained environments.

RANK_REASON Academic paper detailing a benchmark comparison of existing models. [lever_c_demoted from research: ic=1 ai=1.0]

Read on arXiv cs.AI →

• arXiv
• CIFAR-10
• CIFAR-100
• EfficientNet-B0
• EfficientNetV2-S
• ImageNet
• MobileNetV3-Small
• Tiny ImageNet

• paper
• other

AI-generated summary · Google Gemini · from 1 sources. How we write summaries →