I successfully managed to run the training. However, at epoch 10, the following error occurs:
PTBTokenizer tokenized 12953 tokens at 158483.76 tokens per second.
PTBTokenizer tokenized 841 tokens at 14101.83 tokens per second.
PTBTokenizer tokenized 12953 tokens at 152064.25 tokens per second.
PTBTokenizer tokenized 841 tokens at 19575.43 tokens per second.
PTBTokenizer tokenized 12953 tokens at 168350.68 tokens per second.
PTBTokenizer tokenized 841 tokens at 15360.88 tokens per second.
PTBTokenizer tokenized 12699 tokens at 221347.77 tokens per second.
PTBTokenizer tokenized 15915 tokens at 252954.50 tokens per second.
avg. outputs: 5.0
PTBTokenizer tokenized 12699 tokens at 160521.46 tokens per second.
PTBTokenizer tokenized 15915 tokens at 197760.88 tokens per second.
avg. outputs: 5.0
[[0.12857833550209574, 0.12857833550209574, 0.12857833550209574], [0.12857833550209574, 0.12857833550209574, 0.12857833550209574]]
Traceback (most recent call last):
File "train.py", line 372, in <module>
train(opt)
File "train.py", line 283, in train
eval_score, eval_loss = evaluate(model, criterion, postprocessors, val_loader, result_json_path, logger=logger, alpha=opt.ec_alpha, device=opt.device, debug=opt.debug)
File "/home/data_mnt/AICity-2024-Track2-CVPRW/eval_utils.py", line 237, in evaluate
dvc_eval_version=dvc_eval_version
File "/home/data_mnt/AICity-2024-Track2-CVPRW/eval_utils.py", line 140, in eval_metrics
dvc_score.update(eval_para(dvc_filename, referneces=para_gt_filenames))
File "/home/data_mnt/AICity-2024-Track2-CVPRW/densevid_eval3/eval_para.py", line 13, in eval_para
all_scorer=args.all_scorer)
File "/home/data_mnt/AICity-2024-Track2-CVPRW/densevid_eval3/para_evaluate.py", line 47, in __init__
self.ground_truths = self.import_ground_truths(ground_truth_filenames)
File "/home/data_mnt/AICity-2024-Track2-CVPRW/densevid_eval3/para_evaluate.py", line 89, in import_ground_truths
gt = json.load(open(filename))
FileNotFoundError: [Errno 2] No such file or directory: 'data/captions/bdd_entities_select_val_caption_vehicle.json'
It seems that the bdd_entities_select_val_caption_vehicle.json file is missing from the repository. Is there a separate way to obtain this file?
I successfully managed to run the training. However, at epoch 10, the following error occurs:
It seems that the
bdd_entities_select_val_caption_vehicle.json
file is missing from the repository. Is there a separate way to obtain this file?