metadata

task_categories:
  - token-classification

AutoTrain Dataset for project: bert-political-statements-and-questions-swedish-ner

Dataset Description

This dataset has been automatically processed by AutoTrain for project bert-political-statements-and-questions-swedish-ner.

Languages

The BCP-47 code for the dataset's language is unk.

Dataset Structure

Data Instances

A sample from this dataset looks as follows:

[
  {
    "tokens": [
      "KD",
      "ska",
      "f\u00f6rhandla",
      "v\u00e5rbudget",
      "med",
      "Milj\u00f6partiet"
    ],
    "tags": [
      1,
      6,
      6,
      6,
      6,
      1
    ]
  },
  {
    "tokens": [
      "V\u00e4nsterpartiet",
      "ska",
      "diskutera",
      "h\u00f6stbudget",
      "med",
      "Sverigedemokraterna"
    ],
    "tags": [
      1,
      6,
      6,
      6,
      6,
      1
    ]
  }
]

Dataset Fields

The dataset has the following fields (also called "features"):

{
  "tokens": "Sequence(feature=Value(dtype='string', id=None), length=-1, id=None)",
  "tags": "Sequence(feature=ClassLabel(names=['B-LOC', 'B-ORG', 'B-PER', 'I-LOC', 'I-ORG', 'I-PER', 'UNK'], id=None), length=-1, id=None)"
}

Dataset Splits

This dataset is split into a train and validation split. The split sizes are as follow:

Split name	Num samples
train	3147
valid	821