moroco

References:

• Code
• Huggingface

moroco

Use the following command to load this dataset in TFDS:

ds = tfds.load('huggingface:moroco/moroco')

• Description:

The MOROCO (Moldavian and Romanian Dialectal Corpus) dataset contains 33564 samples of text collected from the news domain.
The samples belong to one of the following six topics:
    - culture
    - finance
    - politics
    - science
    - sports
    - tech

• License: CC BY-SA 4.0 License
• Version: 1.0.0
• Splits:

Split	Examples
'test'	5924
'train'	21719
'validation'	5921

• Features:

{
    "id": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "category": {
        "num_classes": 6,
        "names": [
            "culture",
            "finance",
            "politics",
            "science",
            "sports",
            "tech"
        ],
        "names_file": null,
        "id": null,
        "_type": "ClassLabel"
    },
    "sample": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}