Tao Kong

> I add a mobilenet.py as GETTING_STARTED.md said , and modify the config ,but I can't run the training, how can I train the foveabox with mobilenet,please ? Please provide...

This repo has been achieved. FoveaBox is now supported by mmdetection, and you can check the code here: https://github.com/open-mmlab/mmdetection/tree/master/configs/foveabox

@DuinoDu For the first question, we find that not sharing features could get better detection results. Maybe you can have try about sharing weights with four scales. The objectness prior...

@THTCheng 1. Change the det_cls_score_7/det_cls_score_6/det_cls_score_5/det_cls_score_4‘s layer num_output from 210 to other values. For example, if you have k classes, the value will be 10*k. 2. Change the det_cls_score_reshape_7/det_cls_score_reshape_6/det_cls_score_reshape_5/det_cls_score_reshape_4's second dim...

@Duankaiwwen You should evaluate the detection results according to your datasets, not voc.

I think adding the class-detection module could further boost performance, you can have a comparison.