Training for custom dataset about my_yolo HOT 1 OPEN

Yes.
When u down load this project on u pc.Put u image here : /data/image and u label file on /data/label,also u can change u dir on dataset.py and pre_processing.py
Second ,make sure the category of u data，and build u dataset label files,
1.for example,if your have a image named 111.jpg, build a 111.txt on ./data/lable/,like this ：
Car 785 533 905 644
Car 89 551 291 680
Car 268 546 383 650
Truck 455 522 548 615
Truck 548 522 625 605
Car 1726 484 1919 646
Car 758 557 807 617
Car 633 561 680 597
Car 682 557 718 593
(the meaning of each words,category,x1,y1,x2,y2)
2.def u own one_hot_lable() func in pre_processing.py,make sure every category get a one-hot vecotr
3.change the output of nn in yolo.py.For example,if u got 5 category in u dataset，the output shape must be anchor_box*(5+classs)

from my_yolo.