train.py
Created by: DragonLYL0718
2020-05-24 00:13:24,369-INFO: Test iter 0 2020-05-24 00:13:41,353-INFO: Test iter 100 2020-05-24 00:13:49,655-INFO: Test finish iter 157 2020-05-24 00:13:49,655-INFO: Total number of images: 1253, inference time: 47.22646347602829 fps. 2020-05-24 00:13:49,655-INFO: Start evaluate... Traceback (most recent call last): File "PaddleDetection/tools/train.py", line 366, in main() File "PaddleDetection/tools/train.py", line 288, in main cfg['EvalReader']['dataset']) File "/home/aistudio/PaddleDetection/ppdet/utils/eval_utils.py", line 236, in eval_results map_type=map_type) File "/home/aistudio/PaddleDetection/ppdet/utils/voc_eval.py", line 71, in bbox_eval gt_boxes = t['gt_bbox'][0] KeyError: 'gt_bbox' terminate called without an active exception W0524 00:13:50.482211 5568 init.cc:209] Warning: PaddlePaddle catches a failure signal, it may not work properly W0524 00:13:50.482247 5568 init.cc:211] You could check whether you killed PaddlePaddle thread/process accidentally or report the case to PaddlePaddle W0524 00:13:50.482259 5568 init.cc:214] The detail failure signal is:
W0524 00:13:50.482267 5568 init.cc:217] *** Aborted at 1590250430 (unix time) try "date -d @1590250430" if you are using GNU date *** terminate called recursively 之前用的别人notebook上面的旧版的paddledetection上可以运行,自己重新下了一个训练了一轮就被杀进程了