YOLOX train your data
you need generate data.txt like follow format (per line-> one image).
img_path1 x1,y1,x2,y2,class_id x1,y1,x2,y2,class_id2
x1,y1,x2,y2 is int type and it belong to 0-img_w ,0-img_h, not 0~1 !!!
img_path is abs path ;must be careful the sign ” ” and “,” in data.txt, there was an example:
/home/sal/images/000011.jpg 284,548,458,851,7 256,393,369,608,1
i.step1 , before train,you need change yolox/exp/yolox_base.py follow you need, i add some explain in it. such as change data.txt path in it.
ii.step2 , change train.py params, just as https://github.com/Megvii-BaseDetection/YOLOX.git ,when you have changed , just run : python train.py