From the course: Deep Learning with Python: Convolutional Neural Networks
Join today to access over 24,700 courses taught by industry experts.
MobileNet - Python Tutorial
From the course: Deep Learning with Python: Convolutional Neural Networks
MobileNet
- [Instructor] MobileNet, introduced by Google in 2017, is a lightweight convolutional neural network architecture designed specifically for mobile and embedded devices where processing power and memory are limited. MobileNet introduced the idea of depth-wise separable convolutions, an efficient alternative to standard convolutional operations. Rather than applying a single convolutional filter across all input channels, MobileNet first applies a depth-wise convolution that is one filter per channel, followed by a point-wise convolution that combines channels. This significantly reduces the number of operations and model size while preserving predictive performance. One of the standout benefits of MobileNet is its computational efficiency. A depth-wise separable convolution requires roughly eight to nine times fewer computations than a standard convolution of the same size, which makes MobileNet capable of running in real time on CPUs. For instance, it can perform over 30 frames per…
Download courses and learn on the go
Watch courses on your mobile device without an internet connection. Download courses using your iOS or Android LinkedIn Learning app.