Series Path · 5 parts
Modern CNN Architectures
Efficient and densely-connected CNN designs — DenseNet feature reuse, MobileNet mobile-first efficiency, and MobileNetV4 universal inverted bottlenecks, theory and code.
Curriculum — 5 parts
- 01
DenseNet: Densely Connected Convolutional Networks
DenseNet (Densely Connected Convolutional Networks) represents a significant advancement in deep learning architecture design, introduced by Gao Huang, Zhuang…
7 min read - 02
DenseNet: A Code Guide
DenseNet (Densely Connected Convolutional Networks) represents a paradigm shift in deep learning architecture design, introducing unprecedented connectivity…
13 min read - 03
MobileNet: Efficient Neural Networks for Mobile Vision Applications
MobileNet represents a revolutionary approach to deep learning architecture design, specifically optimized for mobile and embedded vision applications.…
10 min read - 04
Complete MobileNet Code Guide
MobileNet is a family of efficient neural network architectures designed specifically for mobile and embedded devices. The key innovation is the use of…
23 min read - 05
MobileNetV4: A Practitioner’s Deep-Dive
Universal models for the mobile ecosystem (Qin et al., 2024, arXiv:2404.10518v2)
34 min read