Yanxingang

You can run the command in the 'get_coco_dataset.sh' line by line.

git clone https://github.com/pdollar/coco cd coco mkdir images cd images wget -c https://pjreddie.com/media/files/train2014.zip wget -c https://pjreddie.com/media/files/val2014.zip unzip -q train2014.zip unzip -q val2014.zip cd .. wget -c https://pjreddie.com/media/files/instances_train-val2014.zip wget -c https://pjreddie.com/media/files/coco/5k.part wget...

I meet the same situation. The result AP0.5=0.34, which is lower than AP=0.52 of BobLiu's model. But I don't know the reason?

暂时是不支持的，要想用yolov3-tiny，需要自己修改网络

This is loading the weight you have trained.