kor_nli

सन्दर्भ:

मल्टी_एनएलआई

इस डेटासेट को TFDS में लोड करने के लिए निम्नलिखित कमांड का उपयोग करें:

ds = tfds.load('huggingface:kor_nli/multi_nli')
  • विवरण :
Korean Natural  Language Inference datasets
  • लाइसेंस : कोई ज्ञात लाइसेंस नहीं
  • संस्करण : 1.0.0
  • विभाजन :
विभाजित करना उदाहरण
'train' 392702
  • विशेषताएँ :
{
    "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"
    }
}

एस.एन.एल.आई

इस डेटासेट को TFDS में लोड करने के लिए निम्नलिखित कमांड का उपयोग करें:

ds = tfds.load('huggingface:kor_nli/snli')
  • विवरण :
Korean Natural  Language Inference datasets
  • लाइसेंस : कोई ज्ञात लाइसेंस नहीं
  • संस्करण : 1.0.0
  • विभाजन :
विभाजित करना उदाहरण
'train' 550152
  • विशेषताएँ :
{
    "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

इस डेटासेट को TFDS में लोड करने के लिए निम्नलिखित कमांड का उपयोग करें:

ds = tfds.load('huggingface:kor_nli/xnli')
  • विवरण :
Korean Natural  Language Inference datasets
  • लाइसेंस : कोई ज्ञात लाइसेंस नहीं
  • संस्करण : 1.0.0
  • विभाजन :
विभाजित करना उदाहरण
'test' 5010
'validation' 2490
  • विशेषताएँ :
{
    "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"
    }
}