Stay organized with collections Save and categorize content based on your preferences.



Use the following command to load this dataset in TFDS:

ds = tfds.load('huggingface:offenseval2020_tr/offenseval2020-turkish')
  • Description:
OffensEval-TR 2020 is a Turkish offensive language corpus. The corpus consist of randomly sampled tweets, and annotated in a similar way to OffensEval and GermEval.
  • License: No known license
  • Version: 1.0.0
  • Splits:
Split Examples
'test' 3528
'train' 31756
  • Features:
    "id": {
        "dtype": "int32",
        "id": null,
        "_type": "Value"
    "tweet": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    "subtask_a": {
        "num_classes": 2,
        "names": [
        "names_file": null,
        "id": null,
        "_type": "ClassLabel"