Training error
semchan opened this issue · comments
Dear, When run the pgcn_train.py will always pop below error, would you please help me to check what is the problem? thanks a lot.
File "/media/ActionRecognition/PGCN/pgcn_dataset.py", line 355, in _video_centric_sampling
print("self.prop_dict[video.id][0]",self.prop_dict[video.id][0])
KeyError: 'video_validation_0000187'
fixed
ordereddict([('train_ft_path', '/home/datasets/THUMOS14/I3D_video_level/Flow_Val_All'), ('test_ft_path', '/home/datasets/THUMOS14/I3D_video_level/Flow_Test_All'), ('train_dict_path', 'data/thumos14_train_prop_dict.pkl'), ('val_dict_path', 'data/thumos14_val_prop_dict.pkl'), ('test_dict_path', 'data/thumos14_test_prop_dict.pkl'), ('train_prop_file', 'data/bsn_train_proposal_list.txt'), ('test_prop_file', 'data/bsn_test_proposal_list.txt'), ('training_epoch_multiplier', 20), ('testing_epoch_multiplier', 1), ('fg_iou_thresh', 0.7), ('bg_iou_thresh', 0.01), ('incomplete_iou_thresh', 0.3), ('bg_coverage_thresh', 0.02), ('incomplete_overlap_thresh', 0.01), ('prop_per_video', 8), ('fg_ratio', 1), ('bg_ratio', 1), ('incomplete_ratio', 6), ('iou_threshold', 0.7), ('dis_threshold', 0), ('starting_ratio', 0.5), ('ending_ratio', 0.5)])
File parsed. Time:5.93
100%|███████████████████████████████████████████████████████████████████████████████████████████████████████████| 199/199 [00:24<00:00, 8.12it/s]
Dict constructed. Time:24.52
Killed
Could you please more details about the error?
fixed
hello I also meet this problem， could you please give me some advice ？ thanks a lot
You can monitor the usage of CPU memory. This error seems to be related to the usage of memory.
@YangJinnnn I had the same error and it was caused by my typo on the data/dataset_cfg.yaml. Please check your train_ft_path on that file.
Could you please more details about the error?
This is cuased by the first time to save the data into local disk not completely, the second time to run the code will to check the local data, but found there will miss some files. delete all the files which have saved into the local disk, and run the code agian will slove the issue. Or, delete all the code, and re-clone this source code then run, it will also slove the issue. Dear @Alvin-Zeng you can also fix this issue in your next release.^_^ thanks