conll2003_select_n.yaml

# Simple preprocessing pipeline that loads conll2003 and allows to select a subset of documents. Uses all
# documents per default.

# Usage:
# Just use dataset=conll2003_select_n and set any "stop" parameter,
# e.g. dataset.select_n.stop=5 to take only 5 train documents. See _select_n.yaml for further options.

defaults:
  - conll2003_base
  - _select_n
  # for datasets, that do not yet have a validation split, just uncomment the following line
  # to create one from the train split (see _create_validation_split.yaml for further options)
  # - _create_validation_split