search

Layout-Parser / layout-model-training

The scripts for training Detectron2-based Layout Models on popular layout analysis datasets
202 stars 55 forks source link

KeyError: "Dataset 'xxxx-train' is not registered! #8

Closed TITC closed 2 years ago

TITC commented 2 years ago

This error will trigger when I set --num-gpus 2, any advice will be welcome.

  File "xxx/train_net.py", line 162, in main
    trainer = Trainer(cfg)
  File "/root/anaconda3/lib/python3.8/site-packages/detectron2/engine/defaults.py", line 383, in __init__
    data_loader = self.build_train_loader(cfg)
  File "xxxx/layout-model-training/tools/train_net.py", line 68, in build_train_loader
    return build_detection_train_loader(cfg, mapper=mapper)
  File "/root/anaconda3/lib/python3.8/site-packages/detectron2/config/config.py", line 192, in wrapped
    explicit_args = _get_args_from_config(from_config, *args, **kwargs)
  File "/root/anaconda3/lib/python3.8/site-packages/detectron2/config/config.py", line 229, in _get_args_from_config
    ret = from_config_func(*args, **kwargs)
  File "/root/anaconda3/lib/python3.8/site-packages/detectron2/data/build.py", line 310, in _train_loader_from_config
    dataset = get_detection_dataset_dicts(
  File "/root/anaconda3/lib/python3.8/site-packages/detectron2/data/build.py", line 227, in get_detection_dataset_dicts
    dataset_dicts = [DatasetCatalog.get(dataset_name) for dataset_name in names]
  File "/root/anaconda3/lib/python3.8/site-packages/detectron2/data/build.py", line 227, in <listcomp>
    dataset_dicts = [DatasetCatalog.get(dataset_name) for dataset_name in names]
  File "/root/anaconda3/lib/python3.8/site-packages/detectron2/data/catalog.py", line 53, in get
    raise KeyError(
KeyError: "Dataset '(paperTable,)-train' is not registered! Available datasets are: coco_2014_train, coco_2014_val, coco_2014_minival, coco_2014_minival_100, coco_2014_v$
lminusminival, coco_2017_train, coco_2017_val, coco_2017_test, coco_2017_test-dev, coco_2017_val_100, keypoints_coco_2014_train, keypoints_coco_2014_val, keypoints_coco_$
014_minival, keypoints_coco_2014_valminusminival, keypoints_coco_2014_minival_100, keypoints_coco_2017_train, keypoints_coco_2017_val, keypoints_coco_2017_val_100, coco_$
017_train_panoptic_separated, coco_2017_train_panoptic_stuffonly, coco_2017_train_panoptic, coco_2017_val_panoptic_separated, coco_2017_val_panoptic_stuffonly, coco_2017$
val_panoptic, coco_2017_val_100_panoptic_separated, coco_2017_val_100_