在用自己数据集训练mask_rcnn时一直出现问题
Created by: lky-bit
原始数据集采用labelme标注生成 数据集使用X2coco的代码生成 路径:C:\Users\HX\PaddleDetection\cocome --annotations --train --val
报错: (paddle) C:\Users\HX\PaddleDetection>python -u tools/train.py -c configs/mask_rcnn_r50_1x.yml --eval 2020-08-09 13:29:09,650-INFO: If regularizer of a Parameter has been set by 'fluid.ParamAttr' or 'fluid.WeightNormParamAttr' already. The Regularization[L2Decay, regularization_coeff=0.000100] in Optimizer will not take effect, and it will only be applied to other Parameters! 2020-08-09 13:29:10,163-INFO: Dataset C:\Users\HX\PaddleDetection\dataset\railway is not valid for reason above, try searching C:\Users\HX/.cache/paddle/dataset or downloading dataset... Traceback (most recent call last): File "tools/train.py", line 368, in main() File "tools/train.py", line 145, in main eval_reader = create_reader(cfg.EvalReader, devices_num=1) File "C:\Users\HX\PaddleDetection\ppdet\data\reader.py", line 434, in create_reader reader = Reader(**cfg)() File "C:\Users\HX\PaddleDetection\ppdet\data\reader.py", line 208, in init self._roidbs = self._dataset.get_roidb() File "C:\Users\HX\PaddleDetection\ppdet\data\source\dataset.py", line 65, in get_roidb self.image_dir) File "C:\Users\HX\PaddleDetection\ppdet\utils\download.py", line 145, in get_dataset_path format(path, osp.split(path)[-1])) ValueError: Dataset dataset/railway is not valid and cannot parse dataset type 'railway' for automaticly downloading, which only supports 'voc' , 'coco', 'wider_face' and 'fruit' currently
环境:windows10、python3.7、paddledetection