kor_nli

References:

• Code
• Huggingface

multi_nli

Use the following command to load this dataset in TFDS:

ds = tfds.load('huggingface:kor_nli/multi_nli')

• Description:

Korean Natural  Language Inference datasets

• License: No known license
• Version: 1.0.0
• Splits:

• Features:

{
    "premise": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "hypothesis": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "label": {
        "num_classes": 3,
        "names": [
            "entailment",
            "neutral",
            "contradiction"
        ],
        "names_file": null,
        "id": null,
        "_type": "ClassLabel"
    }
}

snli

Use the following command to load this dataset in TFDS:

ds = tfds.load('huggingface:kor_nli/snli')

• Description:

Korean Natural  Language Inference datasets

• License: No known license
• Version: 1.0.0
• Splits:

• Features:

{
    "premise": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "hypothesis": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "label": {
        "num_classes": 3,
        "names": [
            "entailment",
            "neutral",
            "contradiction"
        ],
        "names_file": null,
        "id": null,
        "_type": "ClassLabel"
    }
}

xnli

Use the following command to load this dataset in TFDS:

ds = tfds.load('huggingface:kor_nli/xnli')

• Description:

Korean Natural  Language Inference datasets

• License: No known license
• Version: 1.0.0
• Splits:

• Features:

{
    "premise": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "hypothesis": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "label": {
        "num_classes": 3,
        "names": [
            "entailment",
            "neutral",
            "contradiction"
        ],
        "names_file": null,
        "id": null,
        "_type": "ClassLabel"
    }
}