MobileNet V1 - AshokBhat/ml GitHub Wiki

• Built around smaller, depth-wise-separable convolutions
• Reduced model complexity and computational burden

• Reduces the parameters by 6.1x
• Reduces operations by 6.8x

• MobileNets: Efficient Convolutional Neural Networks for Mobile Vision Applications

We present a class of efficient models called MobileNet for mobile and embedded vision applications. MobileNets are based on a streamlined architecture that uses depth-wise separable convolutions to build lightweight deep neural networks.

We introduce two simple global hyperparameters that efficiently trade-off between latency and accuracy. These hyper-parameters allow the model builder to choose the right sized model for their application based on the constraints of the problem.

We present extensive experiments on resource and accuracy tradeoffs and show strong performance compared to other popular models on ImageNet classification. We then demonstrate the effectiveness of MobileNets across a wide range of applications and use cases including object detection, fine-grain classification, face attributes, and large scale geo-localization

• MobileNet | MobileNet v1 | MobileNet v2
• ResNet-50