hate_speech_filipino

References:

Use the following command to load this dataset in TFDS:

ds = tfds.load('huggingface:hate_speech_filipino')
  • Description:
Contains 10k tweets (training set) that are labeled as hate speech or non-hate speech. Released with 4,232 validation and 4,232 testing samples. Collected during the 2016 Philippine Presidential Elections.
  • License: No known license
  • Version: 1.0.0
  • Splits:
Split Examples
'test' 10000
'train' 10000
'validation' 4232
  • Features:
{
    "text": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "label": {
        "num_classes": 2,
        "names": [
            "0",
            "1"
        ],
        "names_file": null,
        "id": null,
        "_type": "ClassLabel"
    }
}