a repository containing object detection files made with sixgill labeller for testing with facebook detectron and other object counters the idea is to count objects in herds accurately
photo files for the coco dataset are in img json files are coco dataset made on sixgill labelling portal script contains colab python script for counting the objects cow contains test files for testing cattle objects with for now deer contains test pictures for deer objects, similarly horse goat pig sheep repsectivly
once testing is completed then these files will be uploaded and a new data set created to be labelled whih hopefully improve the brain of the AI ML fast RCNN detectron2 model