provide example code how to train Maskrcnn from https://github.com/matterport/Mask_RCNN
1.prepare folder in <train_data_mold> folder -<train_data_mold> is your training data folder with - is your training data picture - is your training data .json file from "labelme" tool generated
2.use json_to_dataset.py create <cv2_mask> <labelme_json> -python json_to_dataset_ok.py train_data_mold/json -it will create <cv2_mask> <labelme_json> in train_data_mold/json -copy <cv2_mask> <labelme_json> in <train_data_mold>
- motify train_mold.py and run to train model -you can comapre train_mold.py and train_bread.py to look what is the difference between them and know how to modify the parameter
4.run test_mold.py --you can comapre comapre test_mold.py and test_bread.py to look what is the difference between them
5.training weights will be generated in folder