MobileNetV2-SSD

This is a caffe implementation of MobileNet-SSD V2 detection network. It is refered to the model released by tensorflow api https://github.com/tensorflow/models/blob/master/research/object_detection/g3doc/detection_model_zoo.md.

You can also refer to this model in the paper Inverted Residuals and Linear Bottlenecks: Mobile Networks for Classification, Detection and Segment, it is Mobilenet V2+SSDLite architecture.

All these file are created for pascal voc which has 21 classes(one is background), you can change it to suit your dataset.

Pretrain model

Caffe model from pascal voc 0712 dataset can be down from(https://pan.baidu.com/s/1Hht2LeFiJsxztrGwU5ZyOg).

test script

mobile_test_save.py is an test file for only two class, you can modify it to suit your class

To do

1. an end to end training script for your own data