no-da-NER / README.md
tollefj's picture
Update README.md
5ead311 verified
metadata
dataset_info:
  features:
    - name: tokens
      sequence: string
    - name: ner_tags
      sequence: int64
    - name: dataset
      dtype: string
  splits:
    - name: train
      num_bytes: 20416089
      num_examples: 94253
    - name: validation
      num_bytes: 4904248
      num_examples: 25864
    - name: test
      num_bytes: 4673139
      num_examples: 25015
  download_size: 8000299
  dataset_size: 29993476
configs:
  - config_name: default
    data_files:
      - split: train
        path: data/train-*
      - split: validation
        path: data/validation-*
      - split: test
        path: data/test-*
task_categories:
  - text-classification
language:
  - 'no'
  - da
size_categories:
  - 10K<n<100K

Created from the three datasets wikiann, norne and dane.

See detailed config below:

dataset_ids = [
    "wikiann",
    "dane",
    "norne", 
]

dataset_subsets = {
    "wikiann": [
        "nn", "no", "da"
    ],
    "dane": [None],
    "norne": ["combined-7"],
}

Unified to the following BIO-scheme:

# O: 0
# B-PER: 1
# I-PER: 2
# B-ORG: 3
# I-ORG: 4
# B-LOC: 5
# I-LOC: 6
# B-MISC: 7
# I-MISC: 8

mappers = {
    "norne": {
        0: 0,
        1: 1,
        2: 2,
        3: 3, 
        4: 4, 
        # PROD->MISC
        5: 7,
        6: 8,
        # LOC -> LOC
        7: 5,
        8: 6,
        # DRV -> MISC (names, but derived)
        9: 7,
        10: 8,
        # EVT -> MISC (events)
        11: 7,
        12: 8,
        # MISC -> MISC
        13: 7,
        14: 8,
    }
}