Yang Zhao

@themathgeek13 Hi, have you got a good result by training from scratch? I train COCO from scratch for 100 epoches, but can only get very low mAP which is about...

@themathgeek13 with eval_coco.py, it is even lower. Following is the results. 

The tconf of noobj is 0

@willie1997 @hsyntemiz @mehameha998 I read the histories of issues, how about your training results? Thanks a lot.

It means datasets with only 2d lables, such as COCO, lsp and so on.

You can send me e-mails. It's not convenient to release my wechat directly.