экстремальный

Использованная литература:

• Код
• Обнимающее лицо

XNLI

Используйте следующую команду, чтобы загрузить этот набор данных в TFDS:

ds = tfds.load('huggingface:xtreme/XNLI')

• Описание :

The Cross-lingual Natural Language Inference (XNLI) corpus is a crowd-sourced collection of 5,000 test and
2,500 dev pairs for the MultiNLI corpus. The pairs are annotated with textual entailment and translated into
14 languages: French, Spanish, German, Greek, Bulgarian, Russian, Turkish, Arabic, Vietnamese, Thai, Chinese,
Hindi, Swahili and Urdu. This results in 112.5k annotated pairs. Each premise can be associated with the
corresponding hypothesis in the 15 languages, summing up to more than 1.5M combinations. The corpus is made to
evaluate how to perform inference in any language (including low-resources ones like Swahili or Urdu) when only
English NLI data is available at training time. One solution is cross-lingual sentence encoding, for which XNLI
is an evaluation benchmark.
The Cross-lingual TRansfer Evaluation of Multilingual Encoders (XTREME) benchmark is a benchmark for the evaluation of
the cross-lingual generalization ability of pre-trained multilingual models. It covers 40 typologically diverse languages
(spanning 12 language families) and includes nine tasks that collectively require reasoning about different levels of
syntax and semantics. The languages in XTREME are selected to maximize language diversity, coverage in existing tasks,
and availability of training data. Among these are many under-studied languages, such as the Dravidian languages Tamil
(spoken in southern India, Sri Lanka, and Singapore), Telugu and Malayalam (spoken mainly in southern India), and the
Niger-Congo languages Swahili and Yoruba, spoken in Africa.

• Лицензия : Нет известной лицензии.
• Версия : 1.0.0
• Расколы :

• Функции :

{
    "language": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "sentence1": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "sentence2": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "gold_label": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

Тайдика

Используйте следующую команду, чтобы загрузить этот набор данных в TFDS:

ds = tfds.load('huggingface:xtreme/tydiqa')

• Описание :

Gold passage task (GoldP): Given a passage that is guaranteed to contain the
             answer, predict the single contiguous span of characters that answers the question. This is more similar to
             existing reading comprehension datasets (as opposed to the information-seeking task outlined above).
             This task is constructed with two goals in mind: (1) more directly comparing with prior work and (2) providing
             a simplified way for researchers to use TyDi QA by providing compatibility with existing code for SQuAD 1.1,
             XQuAD, and MLQA. Toward these goals, the gold passage task differs from the primary task in several ways:
             only the gold answer passage is provided rather than the entire Wikipedia article;
             unanswerable questions have been discarded, similar to MLQA and XQuAD;
             we evaluate with the SQuAD 1.1 metrics like XQuAD; and
            Thai and Japanese are removed since the lack of whitespace breaks some tools.

The Cross-lingual TRansfer Evaluation of Multilingual Encoders (XTREME) benchmark is a benchmark for the evaluation of
the cross-lingual generalization ability of pre-trained multilingual models. It covers 40 typologically diverse languages
(spanning 12 language families) and includes nine tasks that collectively require reasoning about different levels of
syntax and semantics. The languages in XTREME are selected to maximize language diversity, coverage in existing tasks,
and availability of training data. Among these are many under-studied languages, such as the Dravidian languages Tamil
(spoken in southern India, Sri Lanka, and Singapore), Telugu and Malayalam (spoken mainly in southern India), and the
Niger-Congo languages Swahili and Yoruba, spoken in Africa.

• Лицензия : Нет известной лицензии.
• Версия : 1.0.0
• Расколы :

• Функции :

{
    "id": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "title": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "context": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "question": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "answers": {
        "feature": {
            "answer_start": {
                "dtype": "int32",
                "id": null,
                "_type": "Value"
            },
            "text": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            }
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

Отряд

Используйте следующую команду, чтобы загрузить этот набор данных в TFDS:

ds = tfds.load('huggingface:xtreme/SQuAD')

• Описание :

Stanford Question Answering Dataset (SQuAD) is a reading comprehension     dataset, consisting of questions posed by crowdworkers on a set of Wikipedia     articles, where the answer to every question is a segment of text, or span,     from the corresponding reading passage, or the question might be unanswerable.
The Cross-lingual TRansfer Evaluation of Multilingual Encoders (XTREME) benchmark is a benchmark for the evaluation of
the cross-lingual generalization ability of pre-trained multilingual models. It covers 40 typologically diverse languages
(spanning 12 language families) and includes nine tasks that collectively require reasoning about different levels of
syntax and semantics. The languages in XTREME are selected to maximize language diversity, coverage in existing tasks,
and availability of training data. Among these are many under-studied languages, such as the Dravidian languages Tamil
(spoken in southern India, Sri Lanka, and Singapore), Telugu and Malayalam (spoken mainly in southern India), and the
Niger-Congo languages Swahili and Yoruba, spoken in Africa.

• Лицензия : Нет известной лицензии.
• Версия : 1.0.0
• Расколы :

• Функции :

{
    "id": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "title": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "context": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "question": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "answers": {
        "feature": {
            "answer_start": {
                "dtype": "int32",
                "id": null,
                "_type": "Value"
            },
            "text": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            }
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

ПАН-X.af

Используйте следующую команду, чтобы загрузить этот набор данных в TFDS:

ds = tfds.load('huggingface:xtreme/PAN-X.af')

• Описание :

The WikiANN dataset (Pan et al. 2017) is a dataset with NER annotations for PER, ORG and LOC. It has been
constructed using the linked entities in Wikipedia pages for 282 different languages including Danish. The dataset
can be loaded with the DaNLP package:
The Cross-lingual TRansfer Evaluation of Multilingual Encoders (XTREME) benchmark is a benchmark for the evaluation of
the cross-lingual generalization ability of pre-trained multilingual models. It covers 40 typologically diverse languages
(spanning 12 language families) and includes nine tasks that collectively require reasoning about different levels of
syntax and semantics. The languages in XTREME are selected to maximize language diversity, coverage in existing tasks,
and availability of training data. Among these are many under-studied languages, such as the Dravidian languages Tamil
(spoken in southern India, Sri Lanka, and Singapore), Telugu and Malayalam (spoken mainly in southern India), and the
Niger-Congo languages Swahili and Yoruba, spoken in Africa.

• Лицензия : Нет известной лицензии.
• Версия : 1.0.0
• Расколы :

• Функции :

{
    "tokens": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "ner_tags": {
        "feature": {
            "num_classes": 7,
            "names": [
                "O",
                "B-PER",
                "I-PER",
                "B-ORG",
                "I-ORG",
                "B-LOC",
                "I-LOC"
            ],
            "names_file": null,
            "id": null,
            "_type": "ClassLabel"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "langs": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

ПАН-X.ar

Используйте следующую команду, чтобы загрузить этот набор данных в TFDS:

ds = tfds.load('huggingface:xtreme/PAN-X.ar')

• Описание :

The WikiANN dataset (Pan et al. 2017) is a dataset with NER annotations for PER, ORG and LOC. It has been
constructed using the linked entities in Wikipedia pages for 282 different languages including Danish. The dataset
can be loaded with the DaNLP package:
The Cross-lingual TRansfer Evaluation of Multilingual Encoders (XTREME) benchmark is a benchmark for the evaluation of
the cross-lingual generalization ability of pre-trained multilingual models. It covers 40 typologically diverse languages
(spanning 12 language families) and includes nine tasks that collectively require reasoning about different levels of
syntax and semantics. The languages in XTREME are selected to maximize language diversity, coverage in existing tasks,
and availability of training data. Among these are many under-studied languages, such as the Dravidian languages Tamil
(spoken in southern India, Sri Lanka, and Singapore), Telugu and Malayalam (spoken mainly in southern India), and the
Niger-Congo languages Swahili and Yoruba, spoken in Africa.

• Лицензия : Нет известной лицензии.
• Версия : 1.0.0
• Расколы :

• Функции :

{
    "tokens": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "ner_tags": {
        "feature": {
            "num_classes": 7,
            "names": [
                "O",
                "B-PER",
                "I-PER",
                "B-ORG",
                "I-ORG",
                "B-LOC",
                "I-LOC"
            ],
            "names_file": null,
            "id": null,
            "_type": "ClassLabel"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "langs": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

ПАН-X.bg

Используйте следующую команду, чтобы загрузить этот набор данных в TFDS:

ds = tfds.load('huggingface:xtreme/PAN-X.bg')

• Описание :

The WikiANN dataset (Pan et al. 2017) is a dataset with NER annotations for PER, ORG and LOC. It has been
constructed using the linked entities in Wikipedia pages for 282 different languages including Danish. The dataset
can be loaded with the DaNLP package:
The Cross-lingual TRansfer Evaluation of Multilingual Encoders (XTREME) benchmark is a benchmark for the evaluation of
the cross-lingual generalization ability of pre-trained multilingual models. It covers 40 typologically diverse languages
(spanning 12 language families) and includes nine tasks that collectively require reasoning about different levels of
syntax and semantics. The languages in XTREME are selected to maximize language diversity, coverage in existing tasks,
and availability of training data. Among these are many under-studied languages, such as the Dravidian languages Tamil
(spoken in southern India, Sri Lanka, and Singapore), Telugu and Malayalam (spoken mainly in southern India), and the
Niger-Congo languages Swahili and Yoruba, spoken in Africa.

• Лицензия : Нет известной лицензии.
• Версия : 1.0.0
• Расколы :

• Функции :

{
    "tokens": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "ner_tags": {
        "feature": {
            "num_classes": 7,
            "names": [
                "O",
                "B-PER",
                "I-PER",
                "B-ORG",
                "I-ORG",
                "B-LOC",
                "I-LOC"
            ],
            "names_file": null,
            "id": null,
            "_type": "ClassLabel"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "langs": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

ПАН-X.bn

Используйте следующую команду, чтобы загрузить этот набор данных в TFDS:

ds = tfds.load('huggingface:xtreme/PAN-X.bn')

• Описание :

The WikiANN dataset (Pan et al. 2017) is a dataset with NER annotations for PER, ORG and LOC. It has been
constructed using the linked entities in Wikipedia pages for 282 different languages including Danish. The dataset
can be loaded with the DaNLP package:
The Cross-lingual TRansfer Evaluation of Multilingual Encoders (XTREME) benchmark is a benchmark for the evaluation of
the cross-lingual generalization ability of pre-trained multilingual models. It covers 40 typologically diverse languages
(spanning 12 language families) and includes nine tasks that collectively require reasoning about different levels of
syntax and semantics. The languages in XTREME are selected to maximize language diversity, coverage in existing tasks,
and availability of training data. Among these are many under-studied languages, such as the Dravidian languages Tamil
(spoken in southern India, Sri Lanka, and Singapore), Telugu and Malayalam (spoken mainly in southern India), and the
Niger-Congo languages Swahili and Yoruba, spoken in Africa.

• Лицензия : Нет известной лицензии.
• Версия : 1.0.0
• Расколы :

• Функции :

{
    "tokens": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "ner_tags": {
        "feature": {
            "num_classes": 7,
            "names": [
                "O",
                "B-PER",
                "I-PER",
                "B-ORG",
                "I-ORG",
                "B-LOC",
                "I-LOC"
            ],
            "names_file": null,
            "id": null,
            "_type": "ClassLabel"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "langs": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

PAN-X.de

Используйте следующую команду, чтобы загрузить этот набор данных в TFDS:

ds = tfds.load('huggingface:xtreme/PAN-X.de')

• Описание :

The WikiANN dataset (Pan et al. 2017) is a dataset with NER annotations for PER, ORG and LOC. It has been
constructed using the linked entities in Wikipedia pages for 282 different languages including Danish. The dataset
can be loaded with the DaNLP package:
The Cross-lingual TRansfer Evaluation of Multilingual Encoders (XTREME) benchmark is a benchmark for the evaluation of
the cross-lingual generalization ability of pre-trained multilingual models. It covers 40 typologically diverse languages
(spanning 12 language families) and includes nine tasks that collectively require reasoning about different levels of
syntax and semantics. The languages in XTREME are selected to maximize language diversity, coverage in existing tasks,
and availability of training data. Among these are many under-studied languages, such as the Dravidian languages Tamil
(spoken in southern India, Sri Lanka, and Singapore), Telugu and Malayalam (spoken mainly in southern India), and the
Niger-Congo languages Swahili and Yoruba, spoken in Africa.

• Лицензия : Нет известной лицензии.
• Версия : 1.0.0
• Расколы :

• Функции :

{
    "tokens": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "ner_tags": {
        "feature": {
            "num_classes": 7,
            "names": [
                "O",
                "B-PER",
                "I-PER",
                "B-ORG",
                "I-ORG",
                "B-LOC",
                "I-LOC"
            ],
            "names_file": null,
            "id": null,
            "_type": "ClassLabel"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "langs": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

ПАН-X.el

Используйте следующую команду, чтобы загрузить этот набор данных в TFDS:

ds = tfds.load('huggingface:xtreme/PAN-X.el')

• Описание :

The WikiANN dataset (Pan et al. 2017) is a dataset with NER annotations for PER, ORG and LOC. It has been
constructed using the linked entities in Wikipedia pages for 282 different languages including Danish. The dataset
can be loaded with the DaNLP package:
The Cross-lingual TRansfer Evaluation of Multilingual Encoders (XTREME) benchmark is a benchmark for the evaluation of
the cross-lingual generalization ability of pre-trained multilingual models. It covers 40 typologically diverse languages
(spanning 12 language families) and includes nine tasks that collectively require reasoning about different levels of
syntax and semantics. The languages in XTREME are selected to maximize language diversity, coverage in existing tasks,
and availability of training data. Among these are many under-studied languages, such as the Dravidian languages Tamil
(spoken in southern India, Sri Lanka, and Singapore), Telugu and Malayalam (spoken mainly in southern India), and the
Niger-Congo languages Swahili and Yoruba, spoken in Africa.

• Лицензия : Нет известной лицензии.
• Версия : 1.0.0
• Расколы :

• Функции :

{
    "tokens": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "ner_tags": {
        "feature": {
            "num_classes": 7,
            "names": [
                "O",
                "B-PER",
                "I-PER",
                "B-ORG",
                "I-ORG",
                "B-LOC",
                "I-LOC"
            ],
            "names_file": null,
            "id": null,
            "_type": "ClassLabel"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "langs": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

PAN-X.en

Используйте следующую команду, чтобы загрузить этот набор данных в TFDS:

ds = tfds.load('huggingface:xtreme/PAN-X.en')

• Описание :

The WikiANN dataset (Pan et al. 2017) is a dataset with NER annotations for PER, ORG and LOC. It has been
constructed using the linked entities in Wikipedia pages for 282 different languages including Danish. The dataset
can be loaded with the DaNLP package:
The Cross-lingual TRansfer Evaluation of Multilingual Encoders (XTREME) benchmark is a benchmark for the evaluation of
the cross-lingual generalization ability of pre-trained multilingual models. It covers 40 typologically diverse languages
(spanning 12 language families) and includes nine tasks that collectively require reasoning about different levels of
syntax and semantics. The languages in XTREME are selected to maximize language diversity, coverage in existing tasks,
and availability of training data. Among these are many under-studied languages, such as the Dravidian languages Tamil
(spoken in southern India, Sri Lanka, and Singapore), Telugu and Malayalam (spoken mainly in southern India), and the
Niger-Congo languages Swahili and Yoruba, spoken in Africa.

• Лицензия : Нет известной лицензии.
• Версия : 1.0.0
• Расколы :

• Функции :

{
    "tokens": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "ner_tags": {
        "feature": {
            "num_classes": 7,
            "names": [
                "O",
                "B-PER",
                "I-PER",
                "B-ORG",
                "I-ORG",
                "B-LOC",
                "I-LOC"
            ],
            "names_file": null,
            "id": null,
            "_type": "ClassLabel"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "langs": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

PAN-X.es

Используйте следующую команду, чтобы загрузить этот набор данных в TFDS:

ds = tfds.load('huggingface:xtreme/PAN-X.es')

• Описание :

The WikiANN dataset (Pan et al. 2017) is a dataset with NER annotations for PER, ORG and LOC. It has been
constructed using the linked entities in Wikipedia pages for 282 different languages including Danish. The dataset
can be loaded with the DaNLP package:
The Cross-lingual TRansfer Evaluation of Multilingual Encoders (XTREME) benchmark is a benchmark for the evaluation of
the cross-lingual generalization ability of pre-trained multilingual models. It covers 40 typologically diverse languages
(spanning 12 language families) and includes nine tasks that collectively require reasoning about different levels of
syntax and semantics. The languages in XTREME are selected to maximize language diversity, coverage in existing tasks,
and availability of training data. Among these are many under-studied languages, such as the Dravidian languages Tamil
(spoken in southern India, Sri Lanka, and Singapore), Telugu and Malayalam (spoken mainly in southern India), and the
Niger-Congo languages Swahili and Yoruba, spoken in Africa.

• Лицензия : Нет известной лицензии.
• Версия : 1.0.0
• Расколы :

• Функции :

{
    "tokens": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "ner_tags": {
        "feature": {
            "num_classes": 7,
            "names": [
                "O",
                "B-PER",
                "I-PER",
                "B-ORG",
                "I-ORG",
                "B-LOC",
                "I-LOC"
            ],
            "names_file": null,
            "id": null,
            "_type": "ClassLabel"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "langs": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

PAN-X.et

Используйте следующую команду, чтобы загрузить этот набор данных в TFDS:

ds = tfds.load('huggingface:xtreme/PAN-X.et')

• Описание :

The WikiANN dataset (Pan et al. 2017) is a dataset with NER annotations for PER, ORG and LOC. It has been
constructed using the linked entities in Wikipedia pages for 282 different languages including Danish. The dataset
can be loaded with the DaNLP package:
The Cross-lingual TRansfer Evaluation of Multilingual Encoders (XTREME) benchmark is a benchmark for the evaluation of
the cross-lingual generalization ability of pre-trained multilingual models. It covers 40 typologically diverse languages
(spanning 12 language families) and includes nine tasks that collectively require reasoning about different levels of
syntax and semantics. The languages in XTREME are selected to maximize language diversity, coverage in existing tasks,
and availability of training data. Among these are many under-studied languages, such as the Dravidian languages Tamil
(spoken in southern India, Sri Lanka, and Singapore), Telugu and Malayalam (spoken mainly in southern India), and the
Niger-Congo languages Swahili and Yoruba, spoken in Africa.

• Лицензия : Нет известной лицензии.
• Версия : 1.0.0
• Расколы :

• Функции :

{
    "tokens": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "ner_tags": {
        "feature": {
            "num_classes": 7,
            "names": [
                "O",
                "B-PER",
                "I-PER",
                "B-ORG",
                "I-ORG",
                "B-LOC",
                "I-LOC"
            ],
            "names_file": null,
            "id": null,
            "_type": "ClassLabel"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "langs": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

PAN-X.eu

Используйте следующую команду, чтобы загрузить этот набор данных в TFDS:

ds = tfds.load('huggingface:xtreme/PAN-X.eu')

• Описание :

The WikiANN dataset (Pan et al. 2017) is a dataset with NER annotations for PER, ORG and LOC. It has been
constructed using the linked entities in Wikipedia pages for 282 different languages including Danish. The dataset
can be loaded with the DaNLP package:
The Cross-lingual TRansfer Evaluation of Multilingual Encoders (XTREME) benchmark is a benchmark for the evaluation of
the cross-lingual generalization ability of pre-trained multilingual models. It covers 40 typologically diverse languages
(spanning 12 language families) and includes nine tasks that collectively require reasoning about different levels of
syntax and semantics. The languages in XTREME are selected to maximize language diversity, coverage in existing tasks,
and availability of training data. Among these are many under-studied languages, such as the Dravidian languages Tamil
(spoken in southern India, Sri Lanka, and Singapore), Telugu and Malayalam (spoken mainly in southern India), and the
Niger-Congo languages Swahili and Yoruba, spoken in Africa.

• Лицензия : Нет известной лицензии.
• Версия : 1.0.0
• Расколы :

• Функции :

{
    "tokens": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "ner_tags": {
        "feature": {
            "num_classes": 7,
            "names": [
                "O",
                "B-PER",
                "I-PER",
                "B-ORG",
                "I-ORG",
                "B-LOC",
                "I-LOC"
            ],
            "names_file": null,
            "id": null,
            "_type": "ClassLabel"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "langs": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

PAN-X.fa

Используйте следующую команду, чтобы загрузить этот набор данных в TFDS:

ds = tfds.load('huggingface:xtreme/PAN-X.fa')

• Описание :

The WikiANN dataset (Pan et al. 2017) is a dataset with NER annotations for PER, ORG and LOC. It has been
constructed using the linked entities in Wikipedia pages for 282 different languages including Danish. The dataset
can be loaded with the DaNLP package:
The Cross-lingual TRansfer Evaluation of Multilingual Encoders (XTREME) benchmark is a benchmark for the evaluation of
the cross-lingual generalization ability of pre-trained multilingual models. It covers 40 typologically diverse languages
(spanning 12 language families) and includes nine tasks that collectively require reasoning about different levels of
syntax and semantics. The languages in XTREME are selected to maximize language diversity, coverage in existing tasks,
and availability of training data. Among these are many under-studied languages, such as the Dravidian languages Tamil
(spoken in southern India, Sri Lanka, and Singapore), Telugu and Malayalam (spoken mainly in southern India), and the
Niger-Congo languages Swahili and Yoruba, spoken in Africa.

• Лицензия : Нет известной лицензии.
• Версия : 1.0.0
• Расколы :

• Функции :

{
    "tokens": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "ner_tags": {
        "feature": {
            "num_classes": 7,
            "names": [
                "O",
                "B-PER",
                "I-PER",
                "B-ORG",
                "I-ORG",
                "B-LOC",
                "I-LOC"
            ],
            "names_file": null,
            "id": null,
            "_type": "ClassLabel"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "langs": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

PAN-X.fi

Используйте следующую команду, чтобы загрузить этот набор данных в TFDS:

ds = tfds.load('huggingface:xtreme/PAN-X.fi')

• Описание :

The WikiANN dataset (Pan et al. 2017) is a dataset with NER annotations for PER, ORG and LOC. It has been
constructed using the linked entities in Wikipedia pages for 282 different languages including Danish. The dataset
can be loaded with the DaNLP package:
The Cross-lingual TRansfer Evaluation of Multilingual Encoders (XTREME) benchmark is a benchmark for the evaluation of
the cross-lingual generalization ability of pre-trained multilingual models. It covers 40 typologically diverse languages
(spanning 12 language families) and includes nine tasks that collectively require reasoning about different levels of
syntax and semantics. The languages in XTREME are selected to maximize language diversity, coverage in existing tasks,
and availability of training data. Among these are many under-studied languages, such as the Dravidian languages Tamil
(spoken in southern India, Sri Lanka, and Singapore), Telugu and Malayalam (spoken mainly in southern India), and the
Niger-Congo languages Swahili and Yoruba, spoken in Africa.

• Лицензия : Нет известной лицензии.
• Версия : 1.0.0
• Расколы :

• Функции :

{
    "tokens": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "ner_tags": {
        "feature": {
            "num_classes": 7,
            "names": [
                "O",
                "B-PER",
                "I-PER",
                "B-ORG",
                "I-ORG",
                "B-LOC",
                "I-LOC"
            ],
            "names_file": null,
            "id": null,
            "_type": "ClassLabel"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "langs": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

PAN-X.fr

Используйте следующую команду, чтобы загрузить этот набор данных в TFDS:

ds = tfds.load('huggingface:xtreme/PAN-X.fr')

• Описание :

The WikiANN dataset (Pan et al. 2017) is a dataset with NER annotations for PER, ORG and LOC. It has been
constructed using the linked entities in Wikipedia pages for 282 different languages including Danish. The dataset
can be loaded with the DaNLP package:
The Cross-lingual TRansfer Evaluation of Multilingual Encoders (XTREME) benchmark is a benchmark for the evaluation of
the cross-lingual generalization ability of pre-trained multilingual models. It covers 40 typologically diverse languages
(spanning 12 language families) and includes nine tasks that collectively require reasoning about different levels of
syntax and semantics. The languages in XTREME are selected to maximize language diversity, coverage in existing tasks,
and availability of training data. Among these are many under-studied languages, such as the Dravidian languages Tamil
(spoken in southern India, Sri Lanka, and Singapore), Telugu and Malayalam (spoken mainly in southern India), and the
Niger-Congo languages Swahili and Yoruba, spoken in Africa.

• Лицензия : Нет известной лицензии.
• Версия : 1.0.0
• Расколы :

• Функции :

{
    "tokens": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "ner_tags": {
        "feature": {
            "num_classes": 7,
            "names": [
                "O",
                "B-PER",
                "I-PER",
                "B-ORG",
                "I-ORG",
                "B-LOC",
                "I-LOC"
            ],
            "names_file": null,
            "id": null,
            "_type": "ClassLabel"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "langs": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

ПАН-X.he

Используйте следующую команду, чтобы загрузить этот набор данных в TFDS:

ds = tfds.load('huggingface:xtreme/PAN-X.he')

• Описание :

The WikiANN dataset (Pan et al. 2017) is a dataset with NER annotations for PER, ORG and LOC. It has been
constructed using the linked entities in Wikipedia pages for 282 different languages including Danish. The dataset
can be loaded with the DaNLP package:
The Cross-lingual TRansfer Evaluation of Multilingual Encoders (XTREME) benchmark is a benchmark for the evaluation of
the cross-lingual generalization ability of pre-trained multilingual models. It covers 40 typologically diverse languages
(spanning 12 language families) and includes nine tasks that collectively require reasoning about different levels of
syntax and semantics. The languages in XTREME are selected to maximize language diversity, coverage in existing tasks,
and availability of training data. Among these are many under-studied languages, such as the Dravidian languages Tamil
(spoken in southern India, Sri Lanka, and Singapore), Telugu and Malayalam (spoken mainly in southern India), and the
Niger-Congo languages Swahili and Yoruba, spoken in Africa.

• Лицензия : Нет известной лицензии.
• Версия : 1.0.0
• Расколы :

• Функции :

{
    "tokens": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "ner_tags": {
        "feature": {
            "num_classes": 7,
            "names": [
                "O",
                "B-PER",
                "I-PER",
                "B-ORG",
                "I-ORG",
                "B-LOC",
                "I-LOC"
            ],
            "names_file": null,
            "id": null,
            "_type": "ClassLabel"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "langs": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

ПАН-X.hi

Используйте следующую команду, чтобы загрузить этот набор данных в TFDS:

ds = tfds.load('huggingface:xtreme/PAN-X.hi')

• Описание :

The WikiANN dataset (Pan et al. 2017) is a dataset with NER annotations for PER, ORG and LOC. It has been
constructed using the linked entities in Wikipedia pages for 282 different languages including Danish. The dataset
can be loaded with the DaNLP package:
The Cross-lingual TRansfer Evaluation of Multilingual Encoders (XTREME) benchmark is a benchmark for the evaluation of
the cross-lingual generalization ability of pre-trained multilingual models. It covers 40 typologically diverse languages
(spanning 12 language families) and includes nine tasks that collectively require reasoning about different levels of
syntax and semantics. The languages in XTREME are selected to maximize language diversity, coverage in existing tasks,
and availability of training data. Among these are many under-studied languages, such as the Dravidian languages Tamil
(spoken in southern India, Sri Lanka, and Singapore), Telugu and Malayalam (spoken mainly in southern India), and the
Niger-Congo languages Swahili and Yoruba, spoken in Africa.

• Лицензия : Нет известной лицензии.
• Версия : 1.0.0
• Расколы :

• Функции :

{
    "tokens": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "ner_tags": {
        "feature": {
            "num_classes": 7,
            "names": [
                "O",
                "B-PER",
                "I-PER",
                "B-ORG",
                "I-ORG",
                "B-LOC",
                "I-LOC"
            ],
            "names_file": null,
            "id": null,
            "_type": "ClassLabel"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "langs": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

PAN-X.hu

Используйте следующую команду, чтобы загрузить этот набор данных в TFDS:

ds = tfds.load('huggingface:xtreme/PAN-X.hu')

• Описание :

The WikiANN dataset (Pan et al. 2017) is a dataset with NER annotations for PER, ORG and LOC. It has been
constructed using the linked entities in Wikipedia pages for 282 different languages including Danish. The dataset
can be loaded with the DaNLP package:
The Cross-lingual TRansfer Evaluation of Multilingual Encoders (XTREME) benchmark is a benchmark for the evaluation of
the cross-lingual generalization ability of pre-trained multilingual models. It covers 40 typologically diverse languages
(spanning 12 language families) and includes nine tasks that collectively require reasoning about different levels of
syntax and semantics. The languages in XTREME are selected to maximize language diversity, coverage in existing tasks,
and availability of training data. Among these are many under-studied languages, such as the Dravidian languages Tamil
(spoken in southern India, Sri Lanka, and Singapore), Telugu and Malayalam (spoken mainly in southern India), and the
Niger-Congo languages Swahili and Yoruba, spoken in Africa.

• Лицензия : Нет известной лицензии.
• Версия : 1.0.0
• Расколы :

• Функции :

{
    "tokens": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "ner_tags": {
        "feature": {
            "num_classes": 7,
            "names": [
                "O",
                "B-PER",
                "I-PER",
                "B-ORG",
                "I-ORG",
                "B-LOC",
                "I-LOC"
            ],
            "names_file": null,
            "id": null,
            "_type": "ClassLabel"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "langs": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

PAN-X.id

Используйте следующую команду, чтобы загрузить этот набор данных в TFDS:

ds = tfds.load('huggingface:xtreme/PAN-X.id')

• Описание :

The WikiANN dataset (Pan et al. 2017) is a dataset with NER annotations for PER, ORG and LOC. It has been
constructed using the linked entities in Wikipedia pages for 282 different languages including Danish. The dataset
can be loaded with the DaNLP package:
The Cross-lingual TRansfer Evaluation of Multilingual Encoders (XTREME) benchmark is a benchmark for the evaluation of
the cross-lingual generalization ability of pre-trained multilingual models. It covers 40 typologically diverse languages
(spanning 12 language families) and includes nine tasks that collectively require reasoning about different levels of
syntax and semantics. The languages in XTREME are selected to maximize language diversity, coverage in existing tasks,
and availability of training data. Among these are many under-studied languages, such as the Dravidian languages Tamil
(spoken in southern India, Sri Lanka, and Singapore), Telugu and Malayalam (spoken mainly in southern India), and the
Niger-Congo languages Swahili and Yoruba, spoken in Africa.

• Лицензия : Нет известной лицензии.
• Версия : 1.0.0
• Расколы :

• Функции :

{
    "tokens": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "ner_tags": {
        "feature": {
            "num_classes": 7,
            "names": [
                "O",
                "B-PER",
                "I-PER",
                "B-ORG",
                "I-ORG",
                "B-LOC",
                "I-LOC"
            ],
            "names_file": null,
            "id": null,
            "_type": "ClassLabel"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "langs": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

ПАН-X.it

Используйте следующую команду, чтобы загрузить этот набор данных в TFDS:

ds = tfds.load('huggingface:xtreme/PAN-X.it')

• Описание :

The WikiANN dataset (Pan et al. 2017) is a dataset with NER annotations for PER, ORG and LOC. It has been
constructed using the linked entities in Wikipedia pages for 282 different languages including Danish. The dataset
can be loaded with the DaNLP package:
The Cross-lingual TRansfer Evaluation of Multilingual Encoders (XTREME) benchmark is a benchmark for the evaluation of
the cross-lingual generalization ability of pre-trained multilingual models. It covers 40 typologically diverse languages
(spanning 12 language families) and includes nine tasks that collectively require reasoning about different levels of
syntax and semantics. The languages in XTREME are selected to maximize language diversity, coverage in existing tasks,
and availability of training data. Among these are many under-studied languages, such as the Dravidian languages Tamil
(spoken in southern India, Sri Lanka, and Singapore), Telugu and Malayalam (spoken mainly in southern India), and the
Niger-Congo languages Swahili and Yoruba, spoken in Africa.

• Лицензия : Нет известной лицензии.
• Версия : 1.0.0
• Расколы :

• Функции :

{
    "tokens": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "ner_tags": {
        "feature": {
            "num_classes": 7,
            "names": [
                "O",
                "B-PER",
                "I-PER",
                "B-ORG",
                "I-ORG",
                "B-LOC",
                "I-LOC"
            ],
            "names_file": null,
            "id": null,
            "_type": "ClassLabel"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "langs": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

ПАН-X.ja

Используйте следующую команду, чтобы загрузить этот набор данных в TFDS:

ds = tfds.load('huggingface:xtreme/PAN-X.ja')

• Описание :

The WikiANN dataset (Pan et al. 2017) is a dataset with NER annotations for PER, ORG and LOC. It has been
constructed using the linked entities in Wikipedia pages for 282 different languages including Danish. The dataset
can be loaded with the DaNLP package:
The Cross-lingual TRansfer Evaluation of Multilingual Encoders (XTREME) benchmark is a benchmark for the evaluation of
the cross-lingual generalization ability of pre-trained multilingual models. It covers 40 typologically diverse languages
(spanning 12 language families) and includes nine tasks that collectively require reasoning about different levels of
syntax and semantics. The languages in XTREME are selected to maximize language diversity, coverage in existing tasks,
and availability of training data. Among these are many under-studied languages, such as the Dravidian languages Tamil
(spoken in southern India, Sri Lanka, and Singapore), Telugu and Malayalam (spoken mainly in southern India), and the
Niger-Congo languages Swahili and Yoruba, spoken in Africa.

• Лицензия : Нет известной лицензии.
• Версия : 1.0.0
• Расколы :

• Функции :

{
    "tokens": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "ner_tags": {
        "feature": {
            "num_classes": 7,
            "names": [
                "O",
                "B-PER",
                "I-PER",
                "B-ORG",
                "I-ORG",
                "B-LOC",
                "I-LOC"
            ],
            "names_file": null,
            "id": null,
            "_type": "ClassLabel"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "langs": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

ПАН-X.jv

Используйте следующую команду, чтобы загрузить этот набор данных в TFDS:

ds = tfds.load('huggingface:xtreme/PAN-X.jv')

• Описание :

The WikiANN dataset (Pan et al. 2017) is a dataset with NER annotations for PER, ORG and LOC. It has been
constructed using the linked entities in Wikipedia pages for 282 different languages including Danish. The dataset
can be loaded with the DaNLP package:
The Cross-lingual TRansfer Evaluation of Multilingual Encoders (XTREME) benchmark is a benchmark for the evaluation of
the cross-lingual generalization ability of pre-trained multilingual models. It covers 40 typologically diverse languages
(spanning 12 language families) and includes nine tasks that collectively require reasoning about different levels of
syntax and semantics. The languages in XTREME are selected to maximize language diversity, coverage in existing tasks,
and availability of training data. Among these are many under-studied languages, such as the Dravidian languages Tamil
(spoken in southern India, Sri Lanka, and Singapore), Telugu and Malayalam (spoken mainly in southern India), and the
Niger-Congo languages Swahili and Yoruba, spoken in Africa.

• Лицензия : Нет известной лицензии.
• Версия : 1.0.0
• Расколы :

• Функции :

{
    "tokens": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "ner_tags": {
        "feature": {
            "num_classes": 7,
            "names": [
                "O",
                "B-PER",
                "I-PER",
                "B-ORG",
                "I-ORG",
                "B-LOC",
                "I-LOC"
            ],
            "names_file": null,
            "id": null,
            "_type": "ClassLabel"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "langs": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

ПАН-Х.ка

Используйте следующую команду, чтобы загрузить этот набор данных в TFDS:

ds = tfds.load('huggingface:xtreme/PAN-X.ka')

• Описание :

The WikiANN dataset (Pan et al. 2017) is a dataset with NER annotations for PER, ORG and LOC. It has been
constructed using the linked entities in Wikipedia pages for 282 different languages including Danish. The dataset
can be loaded with the DaNLP package:
The Cross-lingual TRansfer Evaluation of Multilingual Encoders (XTREME) benchmark is a benchmark for the evaluation of
the cross-lingual generalization ability of pre-trained multilingual models. It covers 40 typologically diverse languages
(spanning 12 language families) and includes nine tasks that collectively require reasoning about different levels of
syntax and semantics. The languages in XTREME are selected to maximize language diversity, coverage in existing tasks,
and availability of training data. Among these are many under-studied languages, such as the Dravidian languages Tamil
(spoken in southern India, Sri Lanka, and Singapore), Telugu and Malayalam (spoken mainly in southern India), and the
Niger-Congo languages Swahili and Yoruba, spoken in Africa.

• Лицензия : Нет известной лицензии.
• Версия : 1.0.0
• Расколы :

• Функции :

{
    "tokens": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "ner_tags": {
        "feature": {
            "num_classes": 7,
            "names": [
                "O",
                "B-PER",
                "I-PER",
                "B-ORG",
                "I-ORG",
                "B-LOC",
                "I-LOC"
            ],
            "names_file": null,
            "id": null,
            "_type": "ClassLabel"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "langs": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

ПАН-X.kk

Используйте следующую команду, чтобы загрузить этот набор данных в TFDS:

ds = tfds.load('huggingface:xtreme/PAN-X.kk')

• Описание :

The WikiANN dataset (Pan et al. 2017) is a dataset with NER annotations for PER, ORG and LOC. It has been
constructed using the linked entities in Wikipedia pages for 282 different languages including Danish. The dataset
can be loaded with the DaNLP package:
The Cross-lingual TRansfer Evaluation of Multilingual Encoders (XTREME) benchmark is a benchmark for the evaluation of
the cross-lingual generalization ability of pre-trained multilingual models. It covers 40 typologically diverse languages
(spanning 12 language families) and includes nine tasks that collectively require reasoning about different levels of
syntax and semantics. The languages in XTREME are selected to maximize language diversity, coverage in existing tasks,
and availability of training data. Among these are many under-studied languages, such as the Dravidian languages Tamil
(spoken in southern India, Sri Lanka, and Singapore), Telugu and Malayalam (spoken mainly in southern India), and the
Niger-Congo languages Swahili and Yoruba, spoken in Africa.

• Лицензия : Нет известной лицензии.
• Версия : 1.0.0
• Расколы :

• Функции :

{
    "tokens": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "ner_tags": {
        "feature": {
            "num_classes": 7,
            "names": [
                "O",
                "B-PER",
                "I-PER",
                "B-ORG",
                "I-ORG",
                "B-LOC",
                "I-LOC"
            ],
            "names_file": null,
            "id": null,
            "_type": "ClassLabel"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "langs": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

ПАН-X.ko

Используйте следующую команду, чтобы загрузить этот набор данных в TFDS:

ds = tfds.load('huggingface:xtreme/PAN-X.ko')

• Описание :

The WikiANN dataset (Pan et al. 2017) is a dataset with NER annotations for PER, ORG and LOC. It has been
constructed using the linked entities in Wikipedia pages for 282 different languages including Danish. The dataset
can be loaded with the DaNLP package:
The Cross-lingual TRansfer Evaluation of Multilingual Encoders (XTREME) benchmark is a benchmark for the evaluation of
the cross-lingual generalization ability of pre-trained multilingual models. It covers 40 typologically diverse languages
(spanning 12 language families) and includes nine tasks that collectively require reasoning about different levels of
syntax and semantics. The languages in XTREME are selected to maximize language diversity, coverage in existing tasks,
and availability of training data. Among these are many under-studied languages, such as the Dravidian languages Tamil
(spoken in southern India, Sri Lanka, and Singapore), Telugu and Malayalam (spoken mainly in southern India), and the
Niger-Congo languages Swahili and Yoruba, spoken in Africa.

• Лицензия : Нет известной лицензии.
• Версия : 1.0.0
• Расколы :

• Функции :

{
    "tokens": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "ner_tags": {
        "feature": {
            "num_classes": 7,
            "names": [
                "O",
                "B-PER",
                "I-PER",
                "B-ORG",
                "I-ORG",
                "B-LOC",
                "I-LOC"
            ],
            "names_file": null,
            "id": null,
            "_type": "ClassLabel"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "langs": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

ПАН-X.ml

Используйте следующую команду, чтобы загрузить этот набор данных в TFDS:

ds = tfds.load('huggingface:xtreme/PAN-X.ml')

• Описание :

The WikiANN dataset (Pan et al. 2017) is a dataset with NER annotations for PER, ORG and LOC. It has been
constructed using the linked entities in Wikipedia pages for 282 different languages including Danish. The dataset
can be loaded with the DaNLP package:
The Cross-lingual TRansfer Evaluation of Multilingual Encoders (XTREME) benchmark is a benchmark for the evaluation of
the cross-lingual generalization ability of pre-trained multilingual models. It covers 40 typologically diverse languages
(spanning 12 language families) and includes nine tasks that collectively require reasoning about different levels of
syntax and semantics. The languages in XTREME are selected to maximize language diversity, coverage in existing tasks,
and availability of training data. Among these are many under-studied languages, such as the Dravidian languages Tamil
(spoken in southern India, Sri Lanka, and Singapore), Telugu and Malayalam (spoken mainly in southern India), and the
Niger-Congo languages Swahili and Yoruba, spoken in Africa.

• Лицензия : Нет известной лицензии.
• Версия : 1.0.0
• Расколы :

• Функции :

{
    "tokens": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "ner_tags": {
        "feature": {
            "num_classes": 7,
            "names": [
                "O",
                "B-PER",
                "I-PER",
                "B-ORG",
                "I-ORG",
                "B-LOC",
                "I-LOC"
            ],
            "names_file": null,
            "id": null,
            "_type": "ClassLabel"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "langs": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

ПАН-X.мр

Используйте следующую команду, чтобы загрузить этот набор данных в TFDS:

ds = tfds.load('huggingface:xtreme/PAN-X.mr')

• Описание :

The WikiANN dataset (Pan et al. 2017) is a dataset with NER annotations for PER, ORG and LOC. It has been
constructed using the linked entities in Wikipedia pages for 282 different languages including Danish. The dataset
can be loaded with the DaNLP package:
The Cross-lingual TRansfer Evaluation of Multilingual Encoders (XTREME) benchmark is a benchmark for the evaluation of
the cross-lingual generalization ability of pre-trained multilingual models. It covers 40 typologically diverse languages
(spanning 12 language families) and includes nine tasks that collectively require reasoning about different levels of
syntax and semantics. The languages in XTREME are selected to maximize language diversity, coverage in existing tasks,
and availability of training data. Among these are many under-studied languages, such as the Dravidian languages Tamil
(spoken in southern India, Sri Lanka, and Singapore), Telugu and Malayalam (spoken mainly in southern India), and the
Niger-Congo languages Swahili and Yoruba, spoken in Africa.

• Лицензия : Нет известной лицензии.
• Версия : 1.0.0
• Расколы :

• Функции :

{
    "tokens": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "ner_tags": {
        "feature": {
            "num_classes": 7,
            "names": [
                "O",
                "B-PER",
                "I-PER",
                "B-ORG",
                "I-ORG",
                "B-LOC",
                "I-LOC"
            ],
            "names_file": null,
            "id": null,
            "_type": "ClassLabel"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "langs": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

ПАН-X.мс

Используйте следующую команду, чтобы загрузить этот набор данных в TFDS:

ds = tfds.load('huggingface:xtreme/PAN-X.ms')

• Описание :

The WikiANN dataset (Pan et al. 2017) is a dataset with NER annotations for PER, ORG and LOC. It has been
constructed using the linked entities in Wikipedia pages for 282 different languages including Danish. The dataset
can be loaded with the DaNLP package:
The Cross-lingual TRansfer Evaluation of Multilingual Encoders (XTREME) benchmark is a benchmark for the evaluation of
the cross-lingual generalization ability of pre-trained multilingual models. It covers 40 typologically diverse languages
(spanning 12 language families) and includes nine tasks that collectively require reasoning about different levels of
syntax and semantics. The languages in XTREME are selected to maximize language diversity, coverage in existing tasks,
and availability of training data. Among these are many under-studied languages, such as the Dravidian languages Tamil
(spoken in southern India, Sri Lanka, and Singapore), Telugu and Malayalam (spoken mainly in southern India), and the
Niger-Congo languages Swahili and Yoruba, spoken in Africa.

• Лицензия : Нет известной лицензии.
• Версия : 1.0.0
• Расколы :

• Функции :

{
    "tokens": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "ner_tags": {
        "feature": {
            "num_classes": 7,
            "names": [
                "O",
                "B-PER",
                "I-PER",
                "B-ORG",
                "I-ORG",
                "B-LOC",
                "I-LOC"
            ],
            "names_file": null,
            "id": null,
            "_type": "ClassLabel"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "langs": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

ПАН-X.my

Используйте следующую команду, чтобы загрузить этот набор данных в TFDS:

ds = tfds.load('huggingface:xtreme/PAN-X.my')

• Описание :

The WikiANN dataset (Pan et al. 2017) is a dataset with NER annotations for PER, ORG and LOC. It has been
constructed using the linked entities in Wikipedia pages for 282 different languages including Danish. The dataset
can be loaded with the DaNLP package:
The Cross-lingual TRansfer Evaluation of Multilingual Encoders (XTREME) benchmark is a benchmark for the evaluation of
the cross-lingual generalization ability of pre-trained multilingual models. It covers 40 typologically diverse languages
(spanning 12 language families) and includes nine tasks that collectively require reasoning about different levels of
syntax and semantics. The languages in XTREME are selected to maximize language diversity, coverage in existing tasks,
and availability of training data. Among these are many under-studied languages, such as the Dravidian languages Tamil
(spoken in southern India, Sri Lanka, and Singapore), Telugu and Malayalam (spoken mainly in southern India), and the
Niger-Congo languages Swahili and Yoruba, spoken in Africa.

• Лицензия : Нет известной лицензии.
• Версия : 1.0.0
• Расколы :

• Функции :

{
    "tokens": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "ner_tags": {
        "feature": {
            "num_classes": 7,
            "names": [
                "O",
                "B-PER",
                "I-PER",
                "B-ORG",
                "I-ORG",
                "B-LOC",
                "I-LOC"
            ],
            "names_file": null,
            "id": null,
            "_type": "ClassLabel"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "langs": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

ПАН-X.nl

Используйте следующую команду, чтобы загрузить этот набор данных в TFDS:

ds = tfds.load('huggingface:xtreme/PAN-X.nl')

• Описание :

The WikiANN dataset (Pan et al. 2017) is a dataset with NER annotations for PER, ORG and LOC. It has been
constructed using the linked entities in Wikipedia pages for 282 different languages including Danish. The dataset
can be loaded with the DaNLP package:
The Cross-lingual TRansfer Evaluation of Multilingual Encoders (XTREME) benchmark is a benchmark for the evaluation of
the cross-lingual generalization ability of pre-trained multilingual models. It covers 40 typologically diverse languages
(spanning 12 language families) and includes nine tasks that collectively require reasoning about different levels of
syntax and semantics. The languages in XTREME are selected to maximize language diversity, coverage in existing tasks,
and availability of training data. Among these are many under-studied languages, such as the Dravidian languages Tamil
(spoken in southern India, Sri Lanka, and Singapore), Telugu and Malayalam (spoken mainly in southern India), and the
Niger-Congo languages Swahili and Yoruba, spoken in Africa.

• Лицензия : Нет известной лицензии.
• Версия : 1.0.0
• Расколы :

• Функции :

{
    "tokens": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "ner_tags": {
        "feature": {
            "num_classes": 7,
            "names": [
                "O",
                "B-PER",
                "I-PER",
                "B-ORG",
                "I-ORG",
                "B-LOC",
                "I-LOC"
            ],
            "names_file": null,
            "id": null,
            "_type": "ClassLabel"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "langs": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

ПАН-X.pt

Используйте следующую команду, чтобы загрузить этот набор данных в TFDS:

ds = tfds.load('huggingface:xtreme/PAN-X.pt')

• Описание :

The WikiANN dataset (Pan et al. 2017) is a dataset with NER annotations for PER, ORG and LOC. It has been
constructed using the linked entities in Wikipedia pages for 282 different languages including Danish. The dataset
can be loaded with the DaNLP package:
The Cross-lingual TRansfer Evaluation of Multilingual Encoders (XTREME) benchmark is a benchmark for the evaluation of
the cross-lingual generalization ability of pre-trained multilingual models. It covers 40 typologically diverse languages
(spanning 12 language families) and includes nine tasks that collectively require reasoning about different levels of
syntax and semantics. The languages in XTREME are selected to maximize language diversity, coverage in existing tasks,
and availability of training data. Among these are many under-studied languages, such as the Dravidian languages Tamil
(spoken in southern India, Sri Lanka, and Singapore), Telugu and Malayalam (spoken mainly in southern India), and the
Niger-Congo languages Swahili and Yoruba, spoken in Africa.

• Лицензия : Нет известной лицензии.
• Версия : 1.0.0
• Расколы :

• Функции :

{
    "tokens": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "ner_tags": {
        "feature": {
            "num_classes": 7,
            "names": [
                "O",
                "B-PER",
                "I-PER",
                "B-ORG",
                "I-ORG",
                "B-LOC",
                "I-LOC"
            ],
            "names_file": null,
            "id": null,
            "_type": "ClassLabel"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "langs": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

ПАН-Х.ру

Используйте следующую команду, чтобы загрузить этот набор данных в TFDS:

ds = tfds.load('huggingface:xtreme/PAN-X.ru')

• Описание :

The WikiANN dataset (Pan et al. 2017) is a dataset with NER annotations for PER, ORG and LOC. It has been
constructed using the linked entities in Wikipedia pages for 282 different languages including Danish. The dataset
can be loaded with the DaNLP package:
The Cross-lingual TRansfer Evaluation of Multilingual Encoders (XTREME) benchmark is a benchmark for the evaluation of
the cross-lingual generalization ability of pre-trained multilingual models. It covers 40 typologically diverse languages
(spanning 12 language families) and includes nine tasks that collectively require reasoning about different levels of
syntax and semantics. The languages in XTREME are selected to maximize language diversity, coverage in existing tasks,
and availability of training data. Among these are many under-studied languages, such as the Dravidian languages Tamil
(spoken in southern India, Sri Lanka, and Singapore), Telugu and Malayalam (spoken mainly in southern India), and the
Niger-Congo languages Swahili and Yoruba, spoken in Africa.

• Лицензия : Нет известной лицензии.
• Версия : 1.0.0
• Расколы :

• Функции :

{
    "tokens": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "ner_tags": {
        "feature": {
            "num_classes": 7,
            "names": [
                "O",
                "B-PER",
                "I-PER",
                "B-ORG",
                "I-ORG",
                "B-LOC",
                "I-LOC"
            ],
            "names_file": null,
            "id": null,
            "_type": "ClassLabel"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "langs": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

ПАН-X.sw

Используйте следующую команду, чтобы загрузить этот набор данных в TFDS:

ds = tfds.load('huggingface:xtreme/PAN-X.sw')

• Описание :

The WikiANN dataset (Pan et al. 2017) is a dataset with NER annotations for PER, ORG and LOC. It has been
constructed using the linked entities in Wikipedia pages for 282 different languages including Danish. The dataset
can be loaded with the DaNLP package:
The Cross-lingual TRansfer Evaluation of Multilingual Encoders (XTREME) benchmark is a benchmark for the evaluation of
the cross-lingual generalization ability of pre-trained multilingual models. It covers 40 typologically diverse languages
(spanning 12 language families) and includes nine tasks that collectively require reasoning about different levels of
syntax and semantics. The languages in XTREME are selected to maximize language diversity, coverage in existing tasks,
and availability of training data. Among these are many under-studied languages, such as the Dravidian languages Tamil
(spoken in southern India, Sri Lanka, and Singapore), Telugu and Malayalam (spoken mainly in southern India), and the
Niger-Congo languages Swahili and Yoruba, spoken in Africa.

• Лицензия : Нет известной лицензии.
• Версия : 1.0.0
• Расколы :

• Функции :

{
    "tokens": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "ner_tags": {
        "feature": {
            "num_classes": 7,
            "names": [
                "O",
                "B-PER",
                "I-PER",
                "B-ORG",
                "I-ORG",
                "B-LOC",
                "I-LOC"
            ],
            "names_file": null,
            "id": null,
            "_type": "ClassLabel"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "langs": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

PAN-X.ta

Используйте следующую команду, чтобы загрузить этот набор данных в TFDS:

ds = tfds.load('huggingface:xtreme/PAN-X.ta')

• Описание :

The WikiANN dataset (Pan et al. 2017) is a dataset with NER annotations for PER, ORG and LOC. It has been
constructed using the linked entities in Wikipedia pages for 282 different languages including Danish. The dataset
can be loaded with the DaNLP package:
The Cross-lingual TRansfer Evaluation of Multilingual Encoders (XTREME) benchmark is a benchmark for the evaluation of
the cross-lingual generalization ability of pre-trained multilingual models. It covers 40 typologically diverse languages
(spanning 12 language families) and includes nine tasks that collectively require reasoning about different levels of
syntax and semantics. The languages in XTREME are selected to maximize language diversity, coverage in existing tasks,
and availability of training data. Among these are many under-studied languages, such as the Dravidian languages Tamil
(spoken in southern India, Sri Lanka, and Singapore), Telugu and Malayalam (spoken mainly in southern India), and the
Niger-Congo languages Swahili and Yoruba, spoken in Africa.

• Лицензия : Нет известной лицензии.
• Версия : 1.0.0
• Расколы :

• Функции :

{
    "tokens": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "ner_tags": {
        "feature": {
            "num_classes": 7,
            "names": [
                "O",
                "B-PER",
                "I-PER",
                "B-ORG",
                "I-ORG",
                "B-LOC",
                "I-LOC"
            ],
            "names_file": null,
            "id": null,
            "_type": "ClassLabel"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "langs": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

PAN-X.te

Используйте следующую команду, чтобы загрузить этот набор данных в TFDS:

ds = tfds.load('huggingface:xtreme/PAN-X.te')

• Описание :

The WikiANN dataset (Pan et al. 2017) is a dataset with NER annotations for PER, ORG and LOC. It has been
constructed using the linked entities in Wikipedia pages for 282 different languages including Danish. The dataset
can be loaded with the DaNLP package:
The Cross-lingual TRansfer Evaluation of Multilingual Encoders (XTREME) benchmark is a benchmark for the evaluation of
the cross-lingual generalization ability of pre-trained multilingual models. It covers 40 typologically diverse languages
(spanning 12 language families) and includes nine tasks that collectively require reasoning about different levels of
syntax and semantics. The languages in XTREME are selected to maximize language diversity, coverage in existing tasks,
and availability of training data. Among these are many under-studied languages, such as the Dravidian languages Tamil
(spoken in southern India, Sri Lanka, and Singapore), Telugu and Malayalam (spoken mainly in southern India), and the
Niger-Congo languages Swahili and Yoruba, spoken in Africa.

• Лицензия : Нет известной лицензии.
• Версия : 1.0.0
• Расколы :

• Функции :

{
    "tokens": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "ner_tags": {
        "feature": {
            "num_classes": 7,
            "names": [
                "O",
                "B-PER",
                "I-PER",
                "B-ORG",
                "I-ORG",
                "B-LOC",
                "I-LOC"
            ],
            "names_file": null,
            "id": null,
            "_type": "ClassLabel"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "langs": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

ПАН-X.th

Используйте следующую команду, чтобы загрузить этот набор данных в TFDS:

ds = tfds.load('huggingface:xtreme/PAN-X.th')

• Описание :

The WikiANN dataset (Pan et al. 2017) is a dataset with NER annotations for PER, ORG and LOC. It has been
constructed using the linked entities in Wikipedia pages for 282 different languages including Danish. The dataset
can be loaded with the DaNLP package:
The Cross-lingual TRansfer Evaluation of Multilingual Encoders (XTREME) benchmark is a benchmark for the evaluation of
the cross-lingual generalization ability of pre-trained multilingual models. It covers 40 typologically diverse languages
(spanning 12 language families) and includes nine tasks that collectively require reasoning about different levels of
syntax and semantics. The languages in XTREME are selected to maximize language diversity, coverage in existing tasks,
and availability of training data. Among these are many under-studied languages, such as the Dravidian languages Tamil
(spoken in southern India, Sri Lanka, and Singapore), Telugu and Malayalam (spoken mainly in southern India), and the
Niger-Congo languages Swahili and Yoruba, spoken in Africa.

• Лицензия : Нет известной лицензии.
• Версия : 1.0.0
• Расколы :

• Функции :

{
    "tokens": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "ner_tags": {
        "feature": {
            "num_classes": 7,
            "names": [
                "O",
                "B-PER",
                "I-PER",
                "B-ORG",
                "I-ORG",
                "B-LOC",
                "I-LOC"
            ],
            "names_file": null,
            "id": null,
            "_type": "ClassLabel"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "langs": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

ПАН-X.tl

Используйте следующую команду, чтобы загрузить этот набор данных в TFDS:

ds = tfds.load('huggingface:xtreme/PAN-X.tl')

• Описание :

The WikiANN dataset (Pan et al. 2017) is a dataset with NER annotations for PER, ORG and LOC. It has been
constructed using the linked entities in Wikipedia pages for 282 different languages including Danish. The dataset
can be loaded with the DaNLP package:
The Cross-lingual TRansfer Evaluation of Multilingual Encoders (XTREME) benchmark is a benchmark for the evaluation of
the cross-lingual generalization ability of pre-trained multilingual models. It covers 40 typologically diverse languages
(spanning 12 language families) and includes nine tasks that collectively require reasoning about different levels of
syntax and semantics. The languages in XTREME are selected to maximize language diversity, coverage in existing tasks,
and availability of training data. Among these are many under-studied languages, such as the Dravidian languages Tamil
(spoken in southern India, Sri Lanka, and Singapore), Telugu and Malayalam (spoken mainly in southern India), and the
Niger-Congo languages Swahili and Yoruba, spoken in Africa.

• Лицензия : Нет известной лицензии.
• Версия : 1.0.0
• Расколы :

• Функции :

{
    "tokens": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "ner_tags": {
        "feature": {
            "num_classes": 7,
            "names": [
                "O",
                "B-PER",
                "I-PER",
                "B-ORG",
                "I-ORG",
                "B-LOC",
                "I-LOC"
            ],
            "names_file": null,
            "id": null,
            "_type": "ClassLabel"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "langs": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

ПАН-X.tr

Используйте следующую команду, чтобы загрузить этот набор данных в TFDS:

ds = tfds.load('huggingface:xtreme/PAN-X.tr')

• Описание :

The WikiANN dataset (Pan et al. 2017) is a dataset with NER annotations for PER, ORG and LOC. It has been
constructed using the linked entities in Wikipedia pages for 282 different languages including Danish. The dataset
can be loaded with the DaNLP package:
The Cross-lingual TRansfer Evaluation of Multilingual Encoders (XTREME) benchmark is a benchmark for the evaluation of
the cross-lingual generalization ability of pre-trained multilingual models. It covers 40 typologically diverse languages
(spanning 12 language families) and includes nine tasks that collectively require reasoning about different levels of
syntax and semantics. The languages in XTREME are selected to maximize language diversity, coverage in existing tasks,
and availability of training data. Among these are many under-studied languages, such as the Dravidian languages Tamil
(spoken in southern India, Sri Lanka, and Singapore), Telugu and Malayalam (spoken mainly in southern India), and the
Niger-Congo languages Swahili and Yoruba, spoken in Africa.

• Лицензия : Нет известной лицензии.
• Версия : 1.0.0
• Расколы :

• Функции :

{
    "tokens": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "ner_tags": {
        "feature": {
            "num_classes": 7,
            "names": [
                "O",
                "B-PER",
                "I-PER",
                "B-ORG",
                "I-ORG",
                "B-LOC",
                "I-LOC"
            ],
            "names_file": null,
            "id": null,
            "_type": "ClassLabel"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "langs": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

ПАН-X.ur

Используйте следующую команду, чтобы загрузить этот набор данных в TFDS:

ds = tfds.load('huggingface:xtreme/PAN-X.ur')

• Описание :

The WikiANN dataset (Pan et al. 2017) is a dataset with NER annotations for PER, ORG and LOC. It has been
constructed using the linked entities in Wikipedia pages for 282 different languages including Danish. The dataset
can be loaded with the DaNLP package:
The Cross-lingual TRansfer Evaluation of Multilingual Encoders (XTREME) benchmark is a benchmark for the evaluation of
the cross-lingual generalization ability of pre-trained multilingual models. It covers 40 typologically diverse languages
(spanning 12 language families) and includes nine tasks that collectively require reasoning about different levels of
syntax and semantics. The languages in XTREME are selected to maximize language diversity, coverage in existing tasks,
and availability of training data. Among these are many under-studied languages, such as the Dravidian languages Tamil
(spoken in southern India, Sri Lanka, and Singapore), Telugu and Malayalam (spoken mainly in southern India), and the
Niger-Congo languages Swahili and Yoruba, spoken in Africa.

• Лицензия : Нет известной лицензии.
• Версия : 1.0.0
• Расколы :

• Функции :

{
    "tokens": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "ner_tags": {
        "feature": {
            "num_classes": 7,
            "names": [
                "O",
                "B-PER",
                "I-PER",
                "B-ORG",
                "I-ORG",
                "B-LOC",
                "I-LOC"
            ],
            "names_file": null,
            "id": null,
            "_type": "ClassLabel"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "langs": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

ПАН-X.vi

Используйте следующую команду, чтобы загрузить этот набор данных в TFDS:

ds = tfds.load('huggingface:xtreme/PAN-X.vi')

• Описание :

The WikiANN dataset (Pan et al. 2017) is a dataset with NER annotations for PER, ORG and LOC. It has been
constructed using the linked entities in Wikipedia pages for 282 different languages including Danish. The dataset
can be loaded with the DaNLP package:
The Cross-lingual TRansfer Evaluation of Multilingual Encoders (XTREME) benchmark is a benchmark for the evaluation of
the cross-lingual generalization ability of pre-trained multilingual models. It covers 40 typologically diverse languages
(spanning 12 language families) and includes nine tasks that collectively require reasoning about different levels of
syntax and semantics. The languages in XTREME are selected to maximize language diversity, coverage in existing tasks,
and availability of training data. Among these are many under-studied languages, such as the Dravidian languages Tamil
(spoken in southern India, Sri Lanka, and Singapore), Telugu and Malayalam (spoken mainly in southern India), and the
Niger-Congo languages Swahili and Yoruba, spoken in Africa.

• Лицензия : Нет известной лицензии.
• Версия : 1.0.0
• Расколы :

• Функции :

{
    "tokens": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "ner_tags": {
        "feature": {
            "num_classes": 7,
            "names": [
                "O",
                "B-PER",
                "I-PER",
                "B-ORG",
                "I-ORG",
                "B-LOC",
                "I-LOC"
            ],
            "names_file": null,
            "id": null,
            "_type": "ClassLabel"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "langs": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

ПАН-X.yo

Используйте следующую команду, чтобы загрузить этот набор данных в TFDS:

ds = tfds.load('huggingface:xtreme/PAN-X.yo')

• Описание :

The WikiANN dataset (Pan et al. 2017) is a dataset with NER annotations for PER, ORG and LOC. It has been
constructed using the linked entities in Wikipedia pages for 282 different languages including Danish. The dataset
can be loaded with the DaNLP package:
The Cross-lingual TRansfer Evaluation of Multilingual Encoders (XTREME) benchmark is a benchmark for the evaluation of
the cross-lingual generalization ability of pre-trained multilingual models. It covers 40 typologically diverse languages
(spanning 12 language families) and includes nine tasks that collectively require reasoning about different levels of
syntax and semantics. The languages in XTREME are selected to maximize language diversity, coverage in existing tasks,
and availability of training data. Among these are many under-studied languages, such as the Dravidian languages Tamil
(spoken in southern India, Sri Lanka, and Singapore), Telugu and Malayalam (spoken mainly in southern India), and the
Niger-Congo languages Swahili and Yoruba, spoken in Africa.

• Лицензия : Нет известной лицензии.
• Версия : 1.0.0
• Расколы :

• Функции :

{
    "tokens": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "ner_tags": {
        "feature": {
            "num_classes": 7,
            "names": [
                "O",
                "B-PER",
                "I-PER",
                "B-ORG",
                "I-ORG",
                "B-LOC",
                "I-LOC"
            ],
            "names_file": null,
            "id": null,
            "_type": "ClassLabel"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "langs": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

ПАН-Х.ж

Используйте следующую команду, чтобы загрузить этот набор данных в TFDS:

ds = tfds.load('huggingface:xtreme/PAN-X.zh')

• Описание :

The WikiANN dataset (Pan et al. 2017) is a dataset with NER annotations for PER, ORG and LOC. It has been
constructed using the linked entities in Wikipedia pages for 282 different languages including Danish. The dataset
can be loaded with the DaNLP package:
The Cross-lingual TRansfer Evaluation of Multilingual Encoders (XTREME) benchmark is a benchmark for the evaluation of
the cross-lingual generalization ability of pre-trained multilingual models. It covers 40 typologically diverse languages
(spanning 12 language families) and includes nine tasks that collectively require reasoning about different levels of
syntax and semantics. The languages in XTREME are selected to maximize language diversity, coverage in existing tasks,
and availability of training data. Among these are many under-studied languages, such as the Dravidian languages Tamil
(spoken in southern India, Sri Lanka, and Singapore), Telugu and Malayalam (spoken mainly in southern India), and the
Niger-Congo languages Swahili and Yoruba, spoken in Africa.

• Лицензия : Нет известной лицензии.
• Версия : 1.0.0
• Расколы :

• Функции :

{
    "tokens": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "ner_tags": {
        "feature": {
            "num_classes": 7,
            "names": [
                "O",
                "B-PER",
                "I-PER",
                "B-ORG",
                "I-ORG",
                "B-LOC",
                "I-LOC"
            ],
            "names_file": null,
            "id": null,
            "_type": "ClassLabel"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "langs": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

MLQA.ar.ar

Используйте следующую команду, чтобы загрузить этот набор данных в TFDS:

ds = tfds.load('huggingface:xtreme/MLQA.ar.ar')

• Описание :

MLQA (MultiLingual Question Answering) is a benchmark dataset for evaluating cross-lingual question answering performance.
MLQA consists of over 5K extractive QA instances (12K in English) in SQuAD format in seven languages - English, Arabic,
German, Spanish, Hindi, Vietnamese and Simplified Chinese. MLQA is highly parallel, with QA instances parallel between
4 different languages on average.
The Cross-lingual TRansfer Evaluation of Multilingual Encoders (XTREME) benchmark is a benchmark for the evaluation of
the cross-lingual generalization ability of pre-trained multilingual models. It covers 40 typologically diverse languages
(spanning 12 language families) and includes nine tasks that collectively require reasoning about different levels of
syntax and semantics. The languages in XTREME are selected to maximize language diversity, coverage in existing tasks,
and availability of training data. Among these are many under-studied languages, such as the Dravidian languages Tamil
(spoken in southern India, Sri Lanka, and Singapore), Telugu and Malayalam (spoken mainly in southern India), and the
Niger-Congo languages Swahili and Yoruba, spoken in Africa.

• Лицензия : Нет известной лицензии.
• Версия : 1.0.0
• Расколы :

• Функции :

{
    "id": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "title": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "context": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "question": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "answers": {
        "feature": {
            "answer_start": {
                "dtype": "int32",
                "id": null,
                "_type": "Value"
            },
            "text": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            }
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

MLQA.ar.de

Используйте следующую команду, чтобы загрузить этот набор данных в TFDS:

ds = tfds.load('huggingface:xtreme/MLQA.ar.de')

• Описание :

MLQA (MultiLingual Question Answering) is a benchmark dataset for evaluating cross-lingual question answering performance.
MLQA consists of over 5K extractive QA instances (12K in English) in SQuAD format in seven languages - English, Arabic,
German, Spanish, Hindi, Vietnamese and Simplified Chinese. MLQA is highly parallel, with QA instances parallel between
4 different languages on average.
The Cross-lingual TRansfer Evaluation of Multilingual Encoders (XTREME) benchmark is a benchmark for the evaluation of
the cross-lingual generalization ability of pre-trained multilingual models. It covers 40 typologically diverse languages
(spanning 12 language families) and includes nine tasks that collectively require reasoning about different levels of
syntax and semantics. The languages in XTREME are selected to maximize language diversity, coverage in existing tasks,
and availability of training data. Among these are many under-studied languages, such as the Dravidian languages Tamil
(spoken in southern India, Sri Lanka, and Singapore), Telugu and Malayalam (spoken mainly in southern India), and the
Niger-Congo languages Swahili and Yoruba, spoken in Africa.

• Лицензия : Нет известной лицензии.
• Версия : 1.0.0
• Расколы :

• Функции :

{
    "id": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "title": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "context": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "question": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "answers": {
        "feature": {
            "answer_start": {
                "dtype": "int32",
                "id": null,
                "_type": "Value"
            },
            "text": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            }
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

MLQA.ar.vi

Используйте следующую команду, чтобы загрузить этот набор данных в TFDS:

ds = tfds.load('huggingface:xtreme/MLQA.ar.vi')

• Описание :

MLQA (MultiLingual Question Answering) is a benchmark dataset for evaluating cross-lingual question answering performance.
MLQA consists of over 5K extractive QA instances (12K in English) in SQuAD format in seven languages - English, Arabic,
German, Spanish, Hindi, Vietnamese and Simplified Chinese. MLQA is highly parallel, with QA instances parallel between
4 different languages on average.
The Cross-lingual TRansfer Evaluation of Multilingual Encoders (XTREME) benchmark is a benchmark for the evaluation of
the cross-lingual generalization ability of pre-trained multilingual models. It covers 40 typologically diverse languages
(spanning 12 language families) and includes nine tasks that collectively require reasoning about different levels of
syntax and semantics. The languages in XTREME are selected to maximize language diversity, coverage in existing tasks,
and availability of training data. Among these are many under-studied languages, such as the Dravidian languages Tamil
(spoken in southern India, Sri Lanka, and Singapore), Telugu and Malayalam (spoken mainly in southern India), and the
Niger-Congo languages Swahili and Yoruba, spoken in Africa.

• Лицензия : Нет известной лицензии.
• Версия : 1.0.0
• Расколы :

• Функции :

{
    "id": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "title": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "context": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "question": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "answers": {
        "feature": {
            "answer_start": {
                "dtype": "int32",
                "id": null,
                "_type": "Value"
            },
            "text": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            }
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

MLQA.ar.zh

Используйте следующую команду, чтобы загрузить этот набор данных в TFDS:

ds = tfds.load('huggingface:xtreme/MLQA.ar.zh')

• Описание :

MLQA (MultiLingual Question Answering) is a benchmark dataset for evaluating cross-lingual question answering performance.
MLQA consists of over 5K extractive QA instances (12K in English) in SQuAD format in seven languages - English, Arabic,
German, Spanish, Hindi, Vietnamese and Simplified Chinese. MLQA is highly parallel, with QA instances parallel between
4 different languages on average.
The Cross-lingual TRansfer Evaluation of Multilingual Encoders (XTREME) benchmark is a benchmark for the evaluation of
the cross-lingual generalization ability of pre-trained multilingual models. It covers 40 typologically diverse languages
(spanning 12 language families) and includes nine tasks that collectively require reasoning about different levels of
syntax and semantics. The languages in XTREME are selected to maximize language diversity, coverage in existing tasks,
and availability of training data. Among these are many under-studied languages, such as the Dravidian languages Tamil
(spoken in southern India, Sri Lanka, and Singapore), Telugu and Malayalam (spoken mainly in southern India), and the
Niger-Congo languages Swahili and Yoruba, spoken in Africa.

• Лицензия : Нет известной лицензии.
• Версия : 1.0.0
• Расколы :

• Функции :

{
    "id": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "title": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "context": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "question": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "answers": {
        "feature": {
            "answer_start": {
                "dtype": "int32",
                "id": null,
                "_type": "Value"
            },
            "text": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            }
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

MLQA.ar.en

Используйте следующую команду, чтобы загрузить этот набор данных в TFDS:

ds = tfds.load('huggingface:xtreme/MLQA.ar.en')

• Описание :

MLQA (MultiLingual Question Answering) is a benchmark dataset for evaluating cross-lingual question answering performance.
MLQA consists of over 5K extractive QA instances (12K in English) in SQuAD format in seven languages - English, Arabic,
German, Spanish, Hindi, Vietnamese and Simplified Chinese. MLQA is highly parallel, with QA instances parallel between
4 different languages on average.
The Cross-lingual TRansfer Evaluation of Multilingual Encoders (XTREME) benchmark is a benchmark for the evaluation of
the cross-lingual generalization ability of pre-trained multilingual models. It covers 40 typologically diverse languages
(spanning 12 language families) and includes nine tasks that collectively require reasoning about different levels of
syntax and semantics. The languages in XTREME are selected to maximize language diversity, coverage in existing tasks,
and availability of training data. Among these are many under-studied languages, such as the Dravidian languages Tamil
(spoken in southern India, Sri Lanka, and Singapore), Telugu and Malayalam (spoken mainly in southern India), and the
Niger-Congo languages Swahili and Yoruba, spoken in Africa.

• Лицензия : Нет известной лицензии.
• Версия : 1.0.0
• Расколы :

• Функции :

{
    "id": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "title": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "context": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "question": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "answers": {
        "feature": {
            "answer_start": {
                "dtype": "int32",
                "id": null,
                "_type": "Value"
            },
            "text": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            }
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

MLQA.ar.es

Используйте следующую команду, чтобы загрузить этот набор данных в TFDS:

ds = tfds.load('huggingface:xtreme/MLQA.ar.es')

• Описание :

MLQA (MultiLingual Question Answering) is a benchmark dataset for evaluating cross-lingual question answering performance.
MLQA consists of over 5K extractive QA instances (12K in English) in SQuAD format in seven languages - English, Arabic,
German, Spanish, Hindi, Vietnamese and Simplified Chinese. MLQA is highly parallel, with QA instances parallel between
4 different languages on average.
The Cross-lingual TRansfer Evaluation of Multilingual Encoders (XTREME) benchmark is a benchmark for the evaluation of
the cross-lingual generalization ability of pre-trained multilingual models. It covers 40 typologically diverse languages
(spanning 12 language families) and includes nine tasks that collectively require reasoning about different levels of
syntax and semantics. The languages in XTREME are selected to maximize language diversity, coverage in existing tasks,
and availability of training data. Among these are many under-studied languages, such as the Dravidian languages Tamil
(spoken in southern India, Sri Lanka, and Singapore), Telugu and Malayalam (spoken mainly in southern India), and the
Niger-Congo languages Swahili and Yoruba, spoken in Africa.

• Лицензия : Нет известной лицензии.
• Версия : 1.0.0
• Расколы :

• Функции :

{
    "id": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "title": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "context": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "question": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "answers": {
        "feature": {
            "answer_start": {
                "dtype": "int32",
                "id": null,
                "_type": "Value"
            },
            "text": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            }
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

MLQA.ar.hi

Используйте следующую команду, чтобы загрузить этот набор данных в TFDS:

ds = tfds.load('huggingface:xtreme/MLQA.ar.hi')

• Описание :

MLQA (MultiLingual Question Answering) is a benchmark dataset for evaluating cross-lingual question answering performance.
MLQA consists of over 5K extractive QA instances (12K in English) in SQuAD format in seven languages - English, Arabic,
German, Spanish, Hindi, Vietnamese and Simplified Chinese. MLQA is highly parallel, with QA instances parallel between
4 different languages on average.
The Cross-lingual TRansfer Evaluation of Multilingual Encoders (XTREME) benchmark is a benchmark for the evaluation of
the cross-lingual generalization ability of pre-trained multilingual models. It covers 40 typologically diverse languages
(spanning 12 language families) and includes nine tasks that collectively require reasoning about different levels of
syntax and semantics. The languages in XTREME are selected to maximize language diversity, coverage in existing tasks,
and availability of training data. Among these are many under-studied languages, such as the Dravidian languages Tamil
(spoken in southern India, Sri Lanka, and Singapore), Telugu and Malayalam (spoken mainly in southern India), and the
Niger-Congo languages Swahili and Yoruba, spoken in Africa.

• Лицензия : Нет известной лицензии.
• Версия : 1.0.0
• Расколы :

• Функции :

{
    "id": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "title": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "context": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "question": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "answers": {
        "feature": {
            "answer_start": {
                "dtype": "int32",
                "id": null,
                "_type": "Value"
            },
            "text": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            }
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

MLQA.de.ar

Используйте следующую команду, чтобы загрузить этот набор данных в TFDS:

ds = tfds.load('huggingface:xtreme/MLQA.de.ar')

• Описание :

MLQA (MultiLingual Question Answering) is a benchmark dataset for evaluating cross-lingual question answering performance.
MLQA consists of over 5K extractive QA instances (12K in English) in SQuAD format in seven languages - English, Arabic,
German, Spanish, Hindi, Vietnamese and Simplified Chinese. MLQA is highly parallel, with QA instances parallel between
4 different languages on average.
The Cross-lingual TRansfer Evaluation of Multilingual Encoders (XTREME) benchmark is a benchmark for the evaluation of
the cross-lingual generalization ability of pre-trained multilingual models. It covers 40 typologically diverse languages
(spanning 12 language families) and includes nine tasks that collectively require reasoning about different levels of
syntax and semantics. The languages in XTREME are selected to maximize language diversity, coverage in existing tasks,
and availability of training data. Among these are many under-studied languages, such as the Dravidian languages Tamil
(spoken in southern India, Sri Lanka, and Singapore), Telugu and Malayalam (spoken mainly in southern India), and the
Niger-Congo languages Swahili and Yoruba, spoken in Africa.

• Лицензия : Нет известной лицензии.
• Версия : 1.0.0
• Расколы :

• Функции :

{
    "id": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "title": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "context": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "question": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "answers": {
        "feature": {
            "answer_start": {
                "dtype": "int32",
                "id": null,
                "_type": "Value"
            },
            "text": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            }
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

MLQA.de.de

Используйте следующую команду, чтобы загрузить этот набор данных в TFDS:

ds = tfds.load('huggingface:xtreme/MLQA.de.de')

• Описание :

MLQA (MultiLingual Question Answering) is a benchmark dataset for evaluating cross-lingual question answering performance.
MLQA consists of over 5K extractive QA instances (12K in English) in SQuAD format in seven languages - English, Arabic,
German, Spanish, Hindi, Vietnamese and Simplified Chinese. MLQA is highly parallel, with QA instances parallel between
4 different languages on average.
The Cross-lingual TRansfer Evaluation of Multilingual Encoders (XTREME) benchmark is a benchmark for the evaluation of
the cross-lingual generalization ability of pre-trained multilingual models. It covers 40 typologically diverse languages
(spanning 12 language families) and includes nine tasks that collectively require reasoning about different levels of
syntax and semantics. The languages in XTREME are selected to maximize language diversity, coverage in existing tasks,
and availability of training data. Among these are many under-studied languages, such as the Dravidian languages Tamil
(spoken in southern India, Sri Lanka, and Singapore), Telugu and Malayalam (spoken mainly in southern India), and the
Niger-Congo languages Swahili and Yoruba, spoken in Africa.

• Лицензия : Нет известной лицензии.
• Версия : 1.0.0
• Расколы :

• Функции :

{
    "id": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "title": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "context": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "question": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "answers": {
        "feature": {
            "answer_start": {
                "dtype": "int32",
                "id": null,
                "_type": "Value"
            },
            "text": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            }
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

MLQA.de.vi

Используйте следующую команду, чтобы загрузить этот набор данных в TFDS:

ds = tfds.load('huggingface:xtreme/MLQA.de.vi')

• Описание :

MLQA (MultiLingual Question Answering) is a benchmark dataset for evaluating cross-lingual question answering performance.
MLQA consists of over 5K extractive QA instances (12K in English) in SQuAD format in seven languages - English, Arabic,
German, Spanish, Hindi, Vietnamese and Simplified Chinese. MLQA is highly parallel, with QA instances parallel between
4 different languages on average.
The Cross-lingual TRansfer Evaluation of Multilingual Encoders (XTREME) benchmark is a benchmark for the evaluation of
the cross-lingual generalization ability of pre-trained multilingual models. It covers 40 typologically diverse languages
(spanning 12 language families) and includes nine tasks that collectively require reasoning about different levels of
syntax and semantics. The languages in XTREME are selected to maximize language diversity, coverage in existing tasks,
and availability of training data. Among these are many under-studied languages, such as the Dravidian languages Tamil
(spoken in southern India, Sri Lanka, and Singapore), Telugu and Malayalam (spoken mainly in southern India), and the
Niger-Congo languages Swahili and Yoruba, spoken in Africa.

• Лицензия : Нет известной лицензии.
• Версия : 1.0.0
• Расколы :

• Функции :

{
    "id": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "title": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "context": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "question": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "answers": {
        "feature": {
            "answer_start": {
                "dtype": "int32",
                "id": null,
                "_type": "Value"
            },
            "text": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            }
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

MLQA.de.zh

Используйте следующую команду, чтобы загрузить этот набор данных в TFDS:

ds = tfds.load('huggingface:xtreme/MLQA.de.zh')

• Описание :

MLQA (MultiLingual Question Answering) is a benchmark dataset for evaluating cross-lingual question answering performance.
MLQA consists of over 5K extractive QA instances (12K in English) in SQuAD format in seven languages - English, Arabic,
German, Spanish, Hindi, Vietnamese and Simplified Chinese. MLQA is highly parallel, with QA instances parallel between
4 different languages on average.
The Cross-lingual TRansfer Evaluation of Multilingual Encoders (XTREME) benchmark is a benchmark for the evaluation of
the cross-lingual generalization ability of pre-trained multilingual models. It covers 40 typologically diverse languages
(spanning 12 language families) and includes nine tasks that collectively require reasoning about different levels of
syntax and semantics. The languages in XTREME are selected to maximize language diversity, coverage in existing tasks,
and availability of training data. Among these are many under-studied languages, such as the Dravidian languages Tamil
(spoken in southern India, Sri Lanka, and Singapore), Telugu and Malayalam (spoken mainly in southern India), and the
Niger-Congo languages Swahili and Yoruba, spoken in Africa.

• Лицензия : Нет известной лицензии.
• Версия : 1.0.0
• Расколы :

• Функции :

{
    "id": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "title": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "context": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "question": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "answers": {
        "feature": {
            "answer_start": {
                "dtype": "int32",
                "id": null,
                "_type": "Value"
            },
            "text": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            }
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

MLQA.de.en

Используйте следующую команду, чтобы загрузить этот набор данных в TFDS:

ds = tfds.load('huggingface:xtreme/MLQA.de.en')

• Описание :

MLQA (MultiLingual Question Answering) is a benchmark dataset for evaluating cross-lingual question answering performance.
MLQA consists of over 5K extractive QA instances (12K in English) in SQuAD format in seven languages - English, Arabic,
German, Spanish, Hindi, Vietnamese and Simplified Chinese. MLQA is highly parallel, with QA instances parallel between
4 different languages on average.
The Cross-lingual TRansfer Evaluation of Multilingual Encoders (XTREME) benchmark is a benchmark for the evaluation of
the cross-lingual generalization ability of pre-trained multilingual models. It covers 40 typologically diverse languages
(spanning 12 language families) and includes nine tasks that collectively require reasoning about different levels of
syntax and semantics. The languages in XTREME are selected to maximize language diversity, coverage in existing tasks,
and availability of training data. Among these are many under-studied languages, such as the Dravidian languages Tamil
(spoken in southern India, Sri Lanka, and Singapore), Telugu and Malayalam (spoken mainly in southern India), and the
Niger-Congo languages Swahili and Yoruba, spoken in Africa.

• Лицензия : Нет известной лицензии.
• Версия : 1.0.0
• Расколы :

• Функции :

{
    "id": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "title": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "context": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "question": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "answers": {
        "feature": {
            "answer_start": {
                "dtype": "int32",
                "id": null,
                "_type": "Value"
            },
            "text": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            }
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

MLQA.de.es

Используйте следующую команду, чтобы загрузить этот набор данных в TFDS:

ds = tfds.load('huggingface:xtreme/MLQA.de.es')

• Описание :

MLQA (MultiLingual Question Answering) is a benchmark dataset for evaluating cross-lingual question answering performance.
MLQA consists of over 5K extractive QA instances (12K in English) in SQuAD format in seven languages - English, Arabic,
German, Spanish, Hindi, Vietnamese and Simplified Chinese. MLQA is highly parallel, with QA instances parallel between
4 different languages on average.
The Cross-lingual TRansfer Evaluation of Multilingual Encoders (XTREME) benchmark is a benchmark for the evaluation of
the cross-lingual generalization ability of pre-trained multilingual models. It covers 40 typologically diverse languages
(spanning 12 language families) and includes nine tasks that collectively require reasoning about different levels of
syntax and semantics. The languages in XTREME are selected to maximize language diversity, coverage in existing tasks,
and availability of training data. Among these are many under-studied languages, such as the Dravidian languages Tamil
(spoken in southern India, Sri Lanka, and Singapore), Telugu and Malayalam (spoken mainly in southern India), and the
Niger-Congo languages Swahili and Yoruba, spoken in Africa.

• Лицензия : Нет известной лицензии.
• Версия : 1.0.0
• Расколы :

• Функции :

{
    "id": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "title": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "context": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "question": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "answers": {
        "feature": {
            "answer_start": {
                "dtype": "int32",
                "id": null,
                "_type": "Value"
            },
            "text": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            }
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

MLQA.de.hi

Используйте следующую команду, чтобы загрузить этот набор данных в TFDS:

ds = tfds.load('huggingface:xtreme/MLQA.de.hi')

• Описание :

MLQA (MultiLingual Question Answering) is a benchmark dataset for evaluating cross-lingual question answering performance.
MLQA consists of over 5K extractive QA instances (12K in English) in SQuAD format in seven languages - English, Arabic,
German, Spanish, Hindi, Vietnamese and Simplified Chinese. MLQA is highly parallel, with QA instances parallel between
4 different languages on average.
The Cross-lingual TRansfer Evaluation of Multilingual Encoders (XTREME) benchmark is a benchmark for the evaluation of
the cross-lingual generalization ability of pre-trained multilingual models. It covers 40 typologically diverse languages
(spanning 12 language families) and includes nine tasks that collectively require reasoning about different levels of
syntax and semantics. The languages in XTREME are selected to maximize language diversity, coverage in existing tasks,
and availability of training data. Among these are many under-studied languages, such as the Dravidian languages Tamil
(spoken in southern India, Sri Lanka, and Singapore), Telugu and Malayalam (spoken mainly in southern India), and the
Niger-Congo languages Swahili and Yoruba, spoken in Africa.

• Лицензия : Нет известной лицензии.
• Версия : 1.0.0
• Расколы :

• Функции :

{
    "id": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "title": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "context": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "question": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "answers": {
        "feature": {
            "answer_start": {
                "dtype": "int32",
                "id": null,
                "_type": "Value"
            },
            "text": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            }
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

MLQA.vi.ar

Используйте следующую команду, чтобы загрузить этот набор данных в TFDS:

ds = tfds.load('huggingface:xtreme/MLQA.vi.ar')

• Описание :

MLQA (MultiLingual Question Answering) is a benchmark dataset for evaluating cross-lingual question answering performance.
MLQA consists of over 5K extractive QA instances (12K in English) in SQuAD format in seven languages - English, Arabic,
German, Spanish, Hindi, Vietnamese and Simplified Chinese. MLQA is highly parallel, with QA instances parallel between
4 different languages on average.
The Cross-lingual TRansfer Evaluation of Multilingual Encoders (XTREME) benchmark is a benchmark for the evaluation of
the cross-lingual generalization ability of pre-trained multilingual models. It covers 40 typologically diverse languages
(spanning 12 language families) and includes nine tasks that collectively require reasoning about different levels of
syntax and semantics. The languages in XTREME are selected to maximize language diversity, coverage in existing tasks,
and availability of training data. Among these are many under-studied languages, such as the Dravidian languages Tamil
(spoken in southern India, Sri Lanka, and Singapore), Telugu and Malayalam (spoken mainly in southern India), and the
Niger-Congo languages Swahili and Yoruba, spoken in Africa.

• Лицензия : Нет известной лицензии.
• Версия : 1.0.0
• Расколы :

• Функции :

{
    "id": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "title": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "context": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "question": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "answers": {
        "feature": {
            "answer_start": {
                "dtype": "int32",
                "id": null,
                "_type": "Value"
            },
            "text": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            }
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

MLQA.vi.de

Используйте следующую команду, чтобы загрузить этот набор данных в TFDS:

ds = tfds.load('huggingface:xtreme/MLQA.vi.de')

• Описание :

MLQA (MultiLingual Question Answering) is a benchmark dataset for evaluating cross-lingual question answering performance.
MLQA consists of over 5K extractive QA instances (12K in English) in SQuAD format in seven languages - English, Arabic,
German, Spanish, Hindi, Vietnamese and Simplified Chinese. MLQA is highly parallel, with QA instances parallel between
4 different languages on average.
The Cross-lingual TRansfer Evaluation of Multilingual Encoders (XTREME) benchmark is a benchmark for the evaluation of
the cross-lingual generalization ability of pre-trained multilingual models. It covers 40 typologically diverse languages
(spanning 12 language families) and includes nine tasks that collectively require reasoning about different levels of
syntax and semantics. The languages in XTREME are selected to maximize language diversity, coverage in existing tasks,
and availability of training data. Among these are many under-studied languages, such as the Dravidian languages Tamil
(spoken in southern India, Sri Lanka, and Singapore), Telugu and Malayalam (spoken mainly in southern India), and the
Niger-Congo languages Swahili and Yoruba, spoken in Africa.

• Лицензия : Нет известной лицензии.
• Версия : 1.0.0
• Расколы :

• Функции :

{
    "id": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "title": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "context": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "question": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "answers": {
        "feature": {
            "answer_start": {
                "dtype": "int32",
                "id": null,
                "_type": "Value"
            },
            "text": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            }
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

MLQA.vi.vi

Используйте следующую команду, чтобы загрузить этот набор данных в TFDS:

ds = tfds.load('huggingface:xtreme/MLQA.vi.vi')

• Описание :

MLQA (MultiLingual Question Answering) is a benchmark dataset for evaluating cross-lingual question answering performance.
MLQA consists of over 5K extractive QA instances (12K in English) in SQuAD format in seven languages - English, Arabic,
German, Spanish, Hindi, Vietnamese and Simplified Chinese. MLQA is highly parallel, with QA instances parallel between
4 different languages on average.
The Cross-lingual TRansfer Evaluation of Multilingual Encoders (XTREME) benchmark is a benchmark for the evaluation of
the cross-lingual generalization ability of pre-trained multilingual models. It covers 40 typologically diverse languages
(spanning 12 language families) and includes nine tasks that collectively require reasoning about different levels of
syntax and semantics. The languages in XTREME are selected to maximize language diversity, coverage in existing tasks,
and availability of training data. Among these are many under-studied languages, such as the Dravidian languages Tamil
(spoken in southern India, Sri Lanka, and Singapore), Telugu and Malayalam (spoken mainly in southern India), and the
Niger-Congo languages Swahili and Yoruba, spoken in Africa.

• Лицензия : Нет известной лицензии.
• Версия : 1.0.0
• Расколы :

• Функции :

{
    "id": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "title": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "context": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "question": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "answers": {
        "feature": {
            "answer_start": {
                "dtype": "int32",
                "id": null,
                "_type": "Value"
            },
            "text": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            }
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

MLQA.vi.zh

Используйте следующую команду, чтобы загрузить этот набор данных в TFDS:

ds = tfds.load('huggingface:xtreme/MLQA.vi.zh')

• Описание :

MLQA (MultiLingual Question Answering) is a benchmark dataset for evaluating cross-lingual question answering performance.
MLQA consists of over 5K extractive QA instances (12K in English) in SQuAD format in seven languages - English, Arabic,
German, Spanish, Hindi, Vietnamese and Simplified Chinese. MLQA is highly parallel, with QA instances parallel between
4 different languages on average.
The Cross-lingual TRansfer Evaluation of Multilingual Encoders (XTREME) benchmark is a benchmark for the evaluation of
the cross-lingual generalization ability of pre-trained multilingual models. It covers 40 typologically diverse languages
(spanning 12 language families) and includes nine tasks that collectively require reasoning about different levels of
syntax and semantics. The languages in XTREME are selected to maximize language diversity, coverage in existing tasks,
and availability of training data. Among these are many under-studied languages, such as the Dravidian languages Tamil
(spoken in southern India, Sri Lanka, and Singapore), Telugu and Malayalam (spoken mainly in southern India), and the
Niger-Congo languages Swahili and Yoruba, spoken in Africa.

• Лицензия : Нет известной лицензии.
• Версия : 1.0.0
• Расколы :

• Функции :

{
    "id": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "title": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "context": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "question": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "answers": {
        "feature": {
            "answer_start": {
                "dtype": "int32",
                "id": null,
                "_type": "Value"
            },
            "text": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            }
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

MLQA.vi.en

Используйте следующую команду, чтобы загрузить этот набор данных в TFDS:

ds = tfds.load('huggingface:xtreme/MLQA.vi.en')

• Описание :

MLQA (MultiLingual Question Answering) is a benchmark dataset for evaluating cross-lingual question answering performance.
MLQA consists of over 5K extractive QA instances (12K in English) in SQuAD format in seven languages - English, Arabic,
German, Spanish, Hindi, Vietnamese and Simplified Chinese. MLQA is highly parallel, with QA instances parallel between
4 different languages on average.
The Cross-lingual TRansfer Evaluation of Multilingual Encoders (XTREME) benchmark is a benchmark for the evaluation of
the cross-lingual generalization ability of pre-trained multilingual models. It covers 40 typologically diverse languages
(spanning 12 language families) and includes nine tasks that collectively require reasoning about different levels of
syntax and semantics. The languages in XTREME are selected to maximize language diversity, coverage in existing tasks,
and availability of training data. Among these are many under-studied languages, such as the Dravidian languages Tamil
(spoken in southern India, Sri Lanka, and Singapore), Telugu and Malayalam (spoken mainly in southern India), and the
Niger-Congo languages Swahili and Yoruba, spoken in Africa.

• Лицензия : Нет известной лицензии.
• Версия : 1.0.0
• Расколы :

• Функции :

{
    "id": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "title": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "context": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "question": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "answers": {
        "feature": {
            "answer_start": {
                "dtype": "int32",
                "id": null,
                "_type": "Value"
            },
            "text": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            }
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

MLQA.vi.es

Используйте следующую команду, чтобы загрузить этот набор данных в TFDS:

ds = tfds.load('huggingface:xtreme/MLQA.vi.es')

• Описание :

MLQA (MultiLingual Question Answering) is a benchmark dataset for evaluating cross-lingual question answering performance.
MLQA consists of over 5K extractive QA instances (12K in English) in SQuAD format in seven languages - English, Arabic,
German, Spanish, Hindi, Vietnamese and Simplified Chinese. MLQA is highly parallel, with QA instances parallel between
4 different languages on average.
The Cross-lingual TRansfer Evaluation of Multilingual Encoders (XTREME) benchmark is a benchmark for the evaluation of
the cross-lingual generalization ability of pre-trained multilingual models. It covers 40 typologically diverse languages
(spanning 12 language families) and includes nine tasks that collectively require reasoning about different levels of
syntax and semantics. The languages in XTREME are selected to maximize language diversity, coverage in existing tasks,
and availability of training data. Among these are many under-studied languages, such as the Dravidian languages Tamil
(spoken in southern India, Sri Lanka, and Singapore), Telugu and Malayalam (spoken mainly in southern India), and the
Niger-Congo languages Swahili and Yoruba, spoken in Africa.

• Лицензия : Нет известной лицензии.
• Версия : 1.0.0
• Расколы :

• Функции :

{
    "id": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "title": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "context": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "question": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "answers": {
        "feature": {
            "answer_start": {
                "dtype": "int32",
                "id": null,
                "_type": "Value"
            },
            "text": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            }
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

MLQA.vi.hi

Используйте следующую команду, чтобы загрузить этот набор данных в TFDS:

ds = tfds.load('huggingface:xtreme/MLQA.vi.hi')

• Описание :

MLQA (MultiLingual Question Answering) is a benchmark dataset for evaluating cross-lingual question answering performance.
MLQA consists of over 5K extractive QA instances (12K in English) in SQuAD format in seven languages - English, Arabic,
German, Spanish, Hindi, Vietnamese and Simplified Chinese. MLQA is highly parallel, with QA instances parallel between
4 different languages on average.
The Cross-lingual TRansfer Evaluation of Multilingual Encoders (XTREME) benchmark is a benchmark for the evaluation of
the cross-lingual generalization ability of pre-trained multilingual models. It covers 40 typologically diverse languages
(spanning 12 language families) and includes nine tasks that collectively require reasoning about different levels of
syntax and semantics. The languages in XTREME are selected to maximize language diversity, coverage in existing tasks,
and availability of training data. Among these are many under-studied languages, such as the Dravidian languages Tamil
(spoken in southern India, Sri Lanka, and Singapore), Telugu and Malayalam (spoken mainly in southern India), and the
Niger-Congo languages Swahili and Yoruba, spoken in Africa.

• Лицензия : Нет известной лицензии.
• Версия : 1.0.0
• Расколы :

• Функции :

{
    "id": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "title": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "context": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "question": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "answers": {
        "feature": {
            "answer_start": {
                "dtype": "int32",
                "id": null,
                "_type": "Value"
            },
            "text": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            }
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

MLQA.ж.ар

Используйте следующую команду, чтобы загрузить этот набор данных в TFDS:

ds = tfds.load('huggingface:xtreme/MLQA.zh.ar')

• Описание :

MLQA (MultiLingual Question Answering) is a benchmark dataset for evaluating cross-lingual question answering performance.
MLQA consists of over 5K extractive QA instances (12K in English) in SQuAD format in seven languages - English, Arabic,
German, Spanish, Hindi, Vietnamese and Simplified Chinese. MLQA is highly parallel, with QA instances parallel between
4 different languages on average.
The Cross-lingual TRansfer Evaluation of Multilingual Encoders (XTREME) benchmark is a benchmark for the evaluation of
the cross-lingual generalization ability of pre-trained multilingual models. It covers 40 typologically diverse languages
(spanning 12 language families) and includes nine tasks that collectively require reasoning about different levels of
syntax and semantics. The languages in XTREME are selected to maximize language diversity, coverage in existing tasks,
and availability of training data. Among these are many under-studied languages, such as the Dravidian languages Tamil
(spoken in southern India, Sri Lanka, and Singapore), Telugu and Malayalam (spoken mainly in southern India), and the
Niger-Congo languages Swahili and Yoruba, spoken in Africa.

• Лицензия : Нет известной лицензии.
• Версия : 1.0.0
• Расколы :

• Функции :

{
    "id": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "title": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "context": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "question": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "answers": {
        "feature": {
            "answer_start": {
                "dtype": "int32",
                "id": null,
                "_type": "Value"
            },
            "text": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            }
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

MLQA.ж.де

Используйте следующую команду, чтобы загрузить этот набор данных в TFDS:

ds = tfds.load('huggingface:xtreme/MLQA.zh.de')

• Описание :

MLQA (MultiLingual Question Answering) is a benchmark dataset for evaluating cross-lingual question answering performance.
MLQA consists of over 5K extractive QA instances (12K in English) in SQuAD format in seven languages - English, Arabic,
German, Spanish, Hindi, Vietnamese and Simplified Chinese. MLQA is highly parallel, with QA instances parallel between
4 different languages on average.
The Cross-lingual TRansfer Evaluation of Multilingual Encoders (XTREME) benchmark is a benchmark for the evaluation of
the cross-lingual generalization ability of pre-trained multilingual models. It covers 40 typologically diverse languages
(spanning 12 language families) and includes nine tasks that collectively require reasoning about different levels of
syntax and semantics. The languages in XTREME are selected to maximize language diversity, coverage in existing tasks,
and availability of training data. Among these are many under-studied languages, such as the Dravidian languages Tamil
(spoken in southern India, Sri Lanka, and Singapore), Telugu and Malayalam (spoken mainly in southern India), and the
Niger-Congo languages Swahili and Yoruba, spoken in Africa.

• Лицензия : Нет известной лицензии.
• Версия : 1.0.0
• Расколы :

• Функции :

{
    "id": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "title": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "context": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "question": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "answers": {
        "feature": {
            "answer_start": {
                "dtype": "int32",
                "id": null,
                "_type": "Value"
            },
            "text": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            }
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

MLQA.zh.vi

Используйте следующую команду, чтобы загрузить этот набор данных в TFDS:

ds = tfds.load('huggingface:xtreme/MLQA.zh.vi')

• Описание :

MLQA (MultiLingual Question Answering) is a benchmark dataset for evaluating cross-lingual question answering performance.
MLQA consists of over 5K extractive QA instances (12K in English) in SQuAD format in seven languages - English, Arabic,
German, Spanish, Hindi, Vietnamese and Simplified Chinese. MLQA is highly parallel, with QA instances parallel between
4 different languages on average.
The Cross-lingual TRansfer Evaluation of Multilingual Encoders (XTREME) benchmark is a benchmark for the evaluation of
the cross-lingual generalization ability of pre-trained multilingual models. It covers 40 typologically diverse languages
(spanning 12 language families) and includes nine tasks that collectively require reasoning about different levels of
syntax and semantics. The languages in XTREME are selected to maximize language diversity, coverage in existing tasks,
and availability of training data. Among these are many under-studied languages, such as the Dravidian languages Tamil
(spoken in southern India, Sri Lanka, and Singapore), Telugu and Malayalam (spoken mainly in southern India), and the
Niger-Congo languages Swahili and Yoruba, spoken in Africa.

• Лицензия : Нет известной лицензии.
• Версия : 1.0.0
• Расколы :

• Функции :

{
    "id": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "title": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "context": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "question": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "answers": {
        "feature": {
            "answer_start": {
                "dtype": "int32",
                "id": null,
                "_type": "Value"
            },
            "text": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            }
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

MLQA.ж.ж

Используйте следующую команду, чтобы загрузить этот набор данных в TFDS:

ds = tfds.load('huggingface:xtreme/MLQA.zh.zh')

• Описание :

MLQA (MultiLingual Question Answering) is a benchmark dataset for evaluating cross-lingual question answering performance.
MLQA consists of over 5K extractive QA instances (12K in English) in SQuAD format in seven languages - English, Arabic,
German, Spanish, Hindi, Vietnamese and Simplified Chinese. MLQA is highly parallel, with QA instances parallel between
4 different languages on average.
The Cross-lingual TRansfer Evaluation of Multilingual Encoders (XTREME) benchmark is a benchmark for the evaluation of
the cross-lingual generalization ability of pre-trained multilingual models. It covers 40 typologically diverse languages
(spanning 12 language families) and includes nine tasks that collectively require reasoning about different levels of
syntax and semantics. The languages in XTREME are selected to maximize language diversity, coverage in existing tasks,
and availability of training data. Among these are many under-studied languages, such as the Dravidian languages Tamil
(spoken in southern India, Sri Lanka, and Singapore), Telugu and Malayalam (spoken mainly in southern India), and the
Niger-Congo languages Swahili and Yoruba, spoken in Africa.

• Лицензия : Нет известной лицензии.
• Версия : 1.0.0
• Расколы :

• Функции :

{
    "id": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "title": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "context": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "question": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "answers": {
        "feature": {
            "answer_start": {
                "dtype": "int32",
                "id": null,
                "_type": "Value"
            },
            "text": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            }
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

MLQA.zh.en

Используйте следующую команду, чтобы загрузить этот набор данных в TFDS:

ds = tfds.load('huggingface:xtreme/MLQA.zh.en')

• Описание :

MLQA (MultiLingual Question Answering) is a benchmark dataset for evaluating cross-lingual question answering performance.
MLQA consists of over 5K extractive QA instances (12K in English) in SQuAD format in seven languages - English, Arabic,
German, Spanish, Hindi, Vietnamese and Simplified Chinese. MLQA is highly parallel, with QA instances parallel between
4 different languages on average.
The Cross-lingual TRansfer Evaluation of Multilingual Encoders (XTREME) benchmark is a benchmark for the evaluation of
the cross-lingual generalization ability of pre-trained multilingual models. It covers 40 typologically diverse languages
(spanning 12 language families) and includes nine tasks that collectively require reasoning about different levels of
syntax and semantics. The languages in XTREME are selected to maximize language diversity, coverage in existing tasks,
and availability of training data. Among these are many under-studied languages, such as the Dravidian languages Tamil
(spoken in southern India, Sri Lanka, and Singapore), Telugu and Malayalam (spoken mainly in southern India), and the
Niger-Congo languages Swahili and Yoruba, spoken in Africa.

• Лицензия : Нет известной лицензии.
• Версия : 1.0.0
• Расколы :

• Функции :

{
    "id": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "title": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "context": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "question": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "answers": {
        "feature": {
            "answer_start": {
                "dtype": "int32",
                "id": null,
                "_type": "Value"
            },
            "text": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            }
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

MLQA.zh.es

Используйте следующую команду, чтобы загрузить этот набор данных в TFDS:

ds = tfds.load('huggingface:xtreme/MLQA.zh.es')

• Описание :

MLQA (MultiLingual Question Answering) is a benchmark dataset for evaluating cross-lingual question answering performance.
MLQA consists of over 5K extractive QA instances (12K in English) in SQuAD format in seven languages - English, Arabic,
German, Spanish, Hindi, Vietnamese and Simplified Chinese. MLQA is highly parallel, with QA instances parallel between
4 different languages on average.
The Cross-lingual TRansfer Evaluation of Multilingual Encoders (XTREME) benchmark is a benchmark for the evaluation of
the cross-lingual generalization ability of pre-trained multilingual models. It covers 40 typologically diverse languages
(spanning 12 language families) and includes nine tasks that collectively require reasoning about different levels of
syntax and semantics. The languages in XTREME are selected to maximize language diversity, coverage in existing tasks,
and availability of training data. Among these are many under-studied languages, such as the Dravidian languages Tamil
(spoken in southern India, Sri Lanka, and Singapore), Telugu and Malayalam (spoken mainly in southern India), and the
Niger-Congo languages Swahili and Yoruba, spoken in Africa.

• Лицензия : Нет известной лицензии.
• Версия : 1.0.0
• Расколы :

• Функции :

{
    "id": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "title": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "context": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "question": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "answers": {
        "feature": {
            "answer_start": {
                "dtype": "int32",
                "id": null,
                "_type": "Value"
            },
            "text": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            }
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

MLQA.ж.хи

Используйте следующую команду, чтобы загрузить этот набор данных в TFDS:

ds = tfds.load('huggingface:xtreme/MLQA.zh.hi')

• Описание :

MLQA (MultiLingual Question Answering) is a benchmark dataset for evaluating cross-lingual question answering performance.
MLQA consists of over 5K extractive QA instances (12K in English) in SQuAD format in seven languages - English, Arabic,
German, Spanish, Hindi, Vietnamese and Simplified Chinese. MLQA is highly parallel, with QA instances parallel between
4 different languages on average.
The Cross-lingual TRansfer Evaluation of Multilingual Encoders (XTREME) benchmark is a benchmark for the evaluation of
the cross-lingual generalization ability of pre-trained multilingual models. It covers 40 typologically diverse languages
(spanning 12 language families) and includes nine tasks that collectively require reasoning about different levels of
syntax and semantics. The languages in XTREME are selected to maximize language diversity, coverage in existing tasks,
and availability of training data. Among these are many under-studied languages, such as the Dravidian languages Tamil
(spoken in southern India, Sri Lanka, and Singapore), Telugu and Malayalam (spoken mainly in southern India), and the
Niger-Congo languages Swahili and Yoruba, spoken in Africa.

• Лицензия : Нет известной лицензии.
• Версия : 1.0.0
• Расколы :

• Функции :

{
    "id": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "title": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "context": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "question": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "answers": {
        "feature": {
            "answer_start": {
                "dtype": "int32",
                "id": null,
                "_type": "Value"
            },
            "text": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            }
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

MLQA.ru.ar

Используйте следующую команду, чтобы загрузить этот набор данных в TFDS:

ds = tfds.load('huggingface:xtreme/MLQA.en.ar')

• Описание :

MLQA (MultiLingual Question Answering) is a benchmark dataset for evaluating cross-lingual question answering performance.
MLQA consists of over 5K extractive QA instances (12K in English) in SQuAD format in seven languages - English, Arabic,
German, Spanish, Hindi, Vietnamese and Simplified Chinese. MLQA is highly parallel, with QA instances parallel between
4 different languages on average.
The Cross-lingual TRansfer Evaluation of Multilingual Encoders (XTREME) benchmark is a benchmark for the evaluation of
the cross-lingual generalization ability of pre-trained multilingual models. It covers 40 typologically diverse languages
(spanning 12 language families) and includes nine tasks that collectively require reasoning about different levels of
syntax and semantics. The languages in XTREME are selected to maximize language diversity, coverage in existing tasks,
and availability of training data. Among these are many under-studied languages, such as the Dravidian languages Tamil
(spoken in southern India, Sri Lanka, and Singapore), Telugu and Malayalam (spoken mainly in southern India), and the
Niger-Congo languages Swahili and Yoruba, spoken in Africa.

• Лицензия : Нет известной лицензии.
• Версия : 1.0.0
• Расколы :

• Функции :

{
    "id": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "title": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "context": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "question": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "answers": {
        "feature": {
            "answer_start": {
                "dtype": "int32",
                "id": null,
                "_type": "Value"
            },
            "text": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            }
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

MLQA.en.de

Используйте следующую команду, чтобы загрузить этот набор данных в TFDS:

ds = tfds.load('huggingface:xtreme/MLQA.en.de')

• Описание :

MLQA (MultiLingual Question Answering) is a benchmark dataset for evaluating cross-lingual question answering performance.
MLQA consists of over 5K extractive QA instances (12K in English) in SQuAD format in seven languages - English, Arabic,
German, Spanish, Hindi, Vietnamese and Simplified Chinese. MLQA is highly parallel, with QA instances parallel between
4 different languages on average.
The Cross-lingual TRansfer Evaluation of Multilingual Encoders (XTREME) benchmark is a benchmark for the evaluation of
the cross-lingual generalization ability of pre-trained multilingual models. It covers 40 typologically diverse languages
(spanning 12 language families) and includes nine tasks that collectively require reasoning about different levels of
syntax and semantics. The languages in XTREME are selected to maximize language diversity, coverage in existing tasks,
and availability of training data. Among these are many under-studied languages, such as the Dravidian languages Tamil
(spoken in southern India, Sri Lanka, and Singapore), Telugu and Malayalam (spoken mainly in southern India), and the
Niger-Congo languages Swahili and Yoruba, spoken in Africa.

• Лицензия : Нет известной лицензии.
• Версия : 1.0.0
• Расколы :

• Функции :

{
    "id": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "title": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "context": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "question": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "answers": {
        "feature": {
            "answer_start": {
                "dtype": "int32",
                "id": null,
                "_type": "Value"
            },
            "text": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            }
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

MLQA.en.vi

Используйте следующую команду, чтобы загрузить этот набор данных в TFDS:

ds = tfds.load('huggingface:xtreme/MLQA.en.vi')

• Описание :

MLQA (MultiLingual Question Answering) is a benchmark dataset for evaluating cross-lingual question answering performance.
MLQA consists of over 5K extractive QA instances (12K in English) in SQuAD format in seven languages - English, Arabic,
German, Spanish, Hindi, Vietnamese and Simplified Chinese. MLQA is highly parallel, with QA instances parallel between
4 different languages on average.
The Cross-lingual TRansfer Evaluation of Multilingual Encoders (XTREME) benchmark is a benchmark for the evaluation of
the cross-lingual generalization ability of pre-trained multilingual models. It covers 40 typologically diverse languages
(spanning 12 language families) and includes nine tasks that collectively require reasoning about different levels of
syntax and semantics. The languages in XTREME are selected to maximize language diversity, coverage in existing tasks,
and availability of training data. Among these are many under-studied languages, such as the Dravidian languages Tamil
(spoken in southern India, Sri Lanka, and Singapore), Telugu and Malayalam (spoken mainly in southern India), and the
Niger-Congo languages Swahili and Yoruba, spoken in Africa.

• Лицензия : Нет известной лицензии.
• Версия : 1.0.0
• Расколы :

• Функции :

{
    "id": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "title": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "context": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "question": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "answers": {
        "feature": {
            "answer_start": {
                "dtype": "int32",
                "id": null,
                "_type": "Value"
            },
            "text": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            }
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

MLQA.en.zh

Используйте следующую команду, чтобы загрузить этот набор данных в TFDS:

ds = tfds.load('huggingface:xtreme/MLQA.en.zh')

• Описание :

MLQA (MultiLingual Question Answering) is a benchmark dataset for evaluating cross-lingual question answering performance.
MLQA consists of over 5K extractive QA instances (12K in English) in SQuAD format in seven languages - English, Arabic,
German, Spanish, Hindi, Vietnamese and Simplified Chinese. MLQA is highly parallel, with QA instances parallel between
4 different languages on average.
The Cross-lingual TRansfer Evaluation of Multilingual Encoders (XTREME) benchmark is a benchmark for the evaluation of
the cross-lingual generalization ability of pre-trained multilingual models. It covers 40 typologically diverse languages
(spanning 12 language families) and includes nine tasks that collectively require reasoning about different levels of
syntax and semantics. The languages in XTREME are selected to maximize language diversity, coverage in existing tasks,
and availability of training data. Among these are many under-studied languages, such as the Dravidian languages Tamil
(spoken in southern India, Sri Lanka, and Singapore), Telugu and Malayalam (spoken mainly in southern India), and the
Niger-Congo languages Swahili and Yoruba, spoken in Africa.

• Лицензия : Нет известной лицензии.
• Версия : 1.0.0
• Расколы :

• Функции :

{
    "id": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "title": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "context": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "question": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "answers": {
        "feature": {
            "answer_start": {
                "dtype": "int32",
                "id": null,
                "_type": "Value"
            },
            "text": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            }
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

MLQA.en.en

Используйте следующую команду, чтобы загрузить этот набор данных в TFDS:

ds = tfds.load('huggingface:xtreme/MLQA.en.en')

• Описание :

MLQA (MultiLingual Question Answering) is a benchmark dataset for evaluating cross-lingual question answering performance.
MLQA consists of over 5K extractive QA instances (12K in English) in SQuAD format in seven languages - English, Arabic,
German, Spanish, Hindi, Vietnamese and Simplified Chinese. MLQA is highly parallel, with QA instances parallel between
4 different languages on average.
The Cross-lingual TRansfer Evaluation of Multilingual Encoders (XTREME) benchmark is a benchmark for the evaluation of
the cross-lingual generalization ability of pre-trained multilingual models. It covers 40 typologically diverse languages
(spanning 12 language families) and includes nine tasks that collectively require reasoning about different levels of
syntax and semantics. The languages in XTREME are selected to maximize language diversity, coverage in existing tasks,
and availability of training data. Among these are many under-studied languages, such as the Dravidian languages Tamil
(spoken in southern India, Sri Lanka, and Singapore), Telugu and Malayalam (spoken mainly in southern India), and the
Niger-Congo languages Swahili and Yoruba, spoken in Africa.

• Лицензия : Нет известной лицензии.
• Версия : 1.0.0
• Расколы :

• Функции :

{
    "id": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "title": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "context": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "question": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "answers": {
        "feature": {
            "answer_start": {
                "dtype": "int32",
                "id": null,
                "_type": "Value"
            },
            "text": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            }
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

MLQA.en.es

Используйте следующую команду, чтобы загрузить этот набор данных в TFDS:

ds = tfds.load('huggingface:xtreme/MLQA.en.es')

• Описание :

MLQA (MultiLingual Question Answering) is a benchmark dataset for evaluating cross-lingual question answering performance.
MLQA consists of over 5K extractive QA instances (12K in English) in SQuAD format in seven languages - English, Arabic,
German, Spanish, Hindi, Vietnamese and Simplified Chinese. MLQA is highly parallel, with QA instances parallel between
4 different languages on average.
The Cross-lingual TRansfer Evaluation of Multilingual Encoders (XTREME) benchmark is a benchmark for the evaluation of
the cross-lingual generalization ability of pre-trained multilingual models. It covers 40 typologically diverse languages
(spanning 12 language families) and includes nine tasks that collectively require reasoning about different levels of
syntax and semantics. The languages in XTREME are selected to maximize language diversity, coverage in existing tasks,
and availability of training data. Among these are many under-studied languages, such as the Dravidian languages Tamil
(spoken in southern India, Sri Lanka, and Singapore), Telugu and Malayalam (spoken mainly in southern India), and the
Niger-Congo languages Swahili and Yoruba, spoken in Africa.

• Лицензия : Нет известной лицензии.
• Версия : 1.0.0
• Расколы :

• Функции :

{
    "id": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "title": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "context": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "question": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "answers": {
        "feature": {
            "answer_start": {
                "dtype": "int32",
                "id": null,
                "_type": "Value"
            },
            "text": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            }
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

MLQA.en.hi

Используйте следующую команду, чтобы загрузить этот набор данных в TFDS:

ds = tfds.load('huggingface:xtreme/MLQA.en.hi')

• Описание :

MLQA (MultiLingual Question Answering) is a benchmark dataset for evaluating cross-lingual question answering performance.
MLQA consists of over 5K extractive QA instances (12K in English) in SQuAD format in seven languages - English, Arabic,
German, Spanish, Hindi, Vietnamese and Simplified Chinese. MLQA is highly parallel, with QA instances parallel between
4 different languages on average.
The Cross-lingual TRansfer Evaluation of Multilingual Encoders (XTREME) benchmark is a benchmark for the evaluation of
the cross-lingual generalization ability of pre-trained multilingual models. It covers 40 typologically diverse languages
(spanning 12 language families) and includes nine tasks that collectively require reasoning about different levels of
syntax and semantics. The languages in XTREME are selected to maximize language diversity, coverage in existing tasks,
and availability of training data. Among these are many under-studied languages, such as the Dravidian languages Tamil
(spoken in southern India, Sri Lanka, and Singapore), Telugu and Malayalam (spoken mainly in southern India), and the
Niger-Congo languages Swahili and Yoruba, spoken in Africa.

• Лицензия : Нет известной лицензии.
• Версия : 1.0.0
• Расколы :

• Функции :

{
    "id": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "title": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "context": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "question": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "answers": {
        "feature": {
            "answer_start": {
                "dtype": "int32",
                "id": null,
                "_type": "Value"
            },
            "text": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            }
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

MLQA.es.ar

Используйте следующую команду, чтобы загрузить этот набор данных в TFDS:

ds = tfds.load('huggingface:xtreme/MLQA.es.ar')

• Описание :

MLQA (MultiLingual Question Answering) is a benchmark dataset for evaluating cross-lingual question answering performance.
MLQA consists of over 5K extractive QA instances (12K in English) in SQuAD format in seven languages - English, Arabic,
German, Spanish, Hindi, Vietnamese and Simplified Chinese. MLQA is highly parallel, with QA instances parallel between
4 different languages on average.
The Cross-lingual TRansfer Evaluation of Multilingual Encoders (XTREME) benchmark is a benchmark for the evaluation of
the cross-lingual generalization ability of pre-trained multilingual models. It covers 40 typologically diverse languages
(spanning 12 language families) and includes nine tasks that collectively require reasoning about different levels of
syntax and semantics. The languages in XTREME are selected to maximize language diversity, coverage in existing tasks,
and availability of training data. Among these are many under-studied languages, such as the Dravidian languages Tamil
(spoken in southern India, Sri Lanka, and Singapore), Telugu and Malayalam (spoken mainly in southern India), and the
Niger-Congo languages Swahili and Yoruba, spoken in Africa.

• Лицензия : Нет известной лицензии.
• Версия : 1.0.0
• Расколы :

• Функции :

{
    "id": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "title": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "context": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "question": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "answers": {
        "feature": {
            "answer_start": {
                "dtype": "int32",
                "id": null,
                "_type": "Value"
            },
            "text": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            }
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

MLQA.es.de

Используйте следующую команду, чтобы загрузить этот набор данных в TFDS:

ds = tfds.load('huggingface:xtreme/MLQA.es.de')

• Описание :

MLQA (MultiLingual Question Answering) is a benchmark dataset for evaluating cross-lingual question answering performance.
MLQA consists of over 5K extractive QA instances (12K in English) in SQuAD format in seven languages - English, Arabic,
German, Spanish, Hindi, Vietnamese and Simplified Chinese. MLQA is highly parallel, with QA instances parallel between
4 different languages on average.
The Cross-lingual TRansfer Evaluation of Multilingual Encoders (XTREME) benchmark is a benchmark for the evaluation of
the cross-lingual generalization ability of pre-trained multilingual models. It covers 40 typologically diverse languages
(spanning 12 language families) and includes nine tasks that collectively require reasoning about different levels of
syntax and semantics. The languages in XTREME are selected to maximize language diversity, coverage in existing tasks,
and availability of training data. Among these are many under-studied languages, such as the Dravidian languages Tamil
(spoken in southern India, Sri Lanka, and Singapore), Telugu and Malayalam (spoken mainly in southern India), and the
Niger-Congo languages Swahili and Yoruba, spoken in Africa.

• Лицензия : Нет известной лицензии.
• Версия : 1.0.0
• Расколы :

• Функции :

{
    "id": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "title": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "context": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "question": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "answers": {
        "feature": {
            "answer_start": {
                "dtype": "int32",
                "id": null,
                "_type": "Value"
            },
            "text": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            }
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

MLQA.es.vi

Используйте следующую команду, чтобы загрузить этот набор данных в TFDS:

ds = tfds.load('huggingface:xtreme/MLQA.es.vi')

• Описание :

MLQA (MultiLingual Question Answering) is a benchmark dataset for evaluating cross-lingual question answering performance.
MLQA consists of over 5K extractive QA instances (12K in English) in SQuAD format in seven languages - English, Arabic,
German, Spanish, Hindi, Vietnamese and Simplified Chinese. MLQA is highly parallel, with QA instances parallel between
4 different languages on average.
The Cross-lingual TRansfer Evaluation of Multilingual Encoders (XTREME) benchmark is a benchmark for the evaluation of
the cross-lingual generalization ability of pre-trained multilingual models. It covers 40 typologically diverse languages
(spanning 12 language families) and includes nine tasks that collectively require reasoning about different levels of
syntax and semantics. The languages in XTREME are selected to maximize language diversity, coverage in existing tasks,
and availability of training data. Among these are many under-studied languages, such as the Dravidian languages Tamil
(spoken in southern India, Sri Lanka, and Singapore), Telugu and Malayalam (spoken mainly in southern India), and the
Niger-Congo languages Swahili and Yoruba, spoken in Africa.

• Лицензия : Нет известной лицензии.
• Версия : 1.0.0
• Расколы :

• Функции :

{
    "id": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "title": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "context": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "question": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "answers": {
        "feature": {
            "answer_start": {
                "dtype": "int32",
                "id": null,
                "_type": "Value"
            },
            "text": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            }
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

MLQA.es.zh

Используйте следующую команду, чтобы загрузить этот набор данных в TFDS:

ds = tfds.load('huggingface:xtreme/MLQA.es.zh')

• Описание :

MLQA (MultiLingual Question Answering) is a benchmark dataset for evaluating cross-lingual question answering performance.
MLQA consists of over 5K extractive QA instances (12K in English) in SQuAD format in seven languages - English, Arabic,
German, Spanish, Hindi, Vietnamese and Simplified Chinese. MLQA is highly parallel, with QA instances parallel between
4 different languages on average.
The Cross-lingual TRansfer Evaluation of Multilingual Encoders (XTREME) benchmark is a benchmark for the evaluation of
the cross-lingual generalization ability of pre-trained multilingual models. It covers 40 typologically diverse languages
(spanning 12 language families) and includes nine tasks that collectively require reasoning about different levels of
syntax and semantics. The languages in XTREME are selected to maximize language diversity, coverage in existing tasks,
and availability of training data. Among these are many under-studied languages, such as the Dravidian languages Tamil
(spoken in southern India, Sri Lanka, and Singapore), Telugu and Malayalam (spoken mainly in southern India), and the
Niger-Congo languages Swahili and Yoruba, spoken in Africa.

• Лицензия : Нет известной лицензии.
• Версия : 1.0.0
• Расколы :

• Функции :

{
    "id": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "title": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "context": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "question": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "answers": {
        "feature": {
            "answer_start": {
                "dtype": "int32",
                "id": null,
                "_type": "Value"
            },
            "text": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            }
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

MLQA.es.en

Используйте следующую команду, чтобы загрузить этот набор данных в TFDS:

ds = tfds.load('huggingface:xtreme/MLQA.es.en')

• Описание :

MLQA (MultiLingual Question Answering) is a benchmark dataset for evaluating cross-lingual question answering performance.
MLQA consists of over 5K extractive QA instances (12K in English) in SQuAD format in seven languages - English, Arabic,
German, Spanish, Hindi, Vietnamese and Simplified Chinese. MLQA is highly parallel, with QA instances parallel between
4 different languages on average.
The Cross-lingual TRansfer Evaluation of Multilingual Encoders (XTREME) benchmark is a benchmark for the evaluation of
the cross-lingual generalization ability of pre-trained multilingual models. It covers 40 typologically diverse languages
(spanning 12 language families) and includes nine tasks that collectively require reasoning about different levels of
syntax and semantics. The languages in XTREME are selected to maximize language diversity, coverage in existing tasks,
and availability of training data. Among these are many under-studied languages, such as the Dravidian languages Tamil
(spoken in southern India, Sri Lanka, and Singapore), Telugu and Malayalam (spoken mainly in southern India), and the
Niger-Congo languages Swahili and Yoruba, spoken in Africa.

• Лицензия : Нет известной лицензии.
• Версия : 1.0.0
• Расколы :

• Функции :

{
    "id": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "title": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "context": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "question": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "answers": {
        "feature": {
            "answer_start": {
                "dtype": "int32",
                "id": null,
                "_type": "Value"
            },
            "text": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            }
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

MLQA.es.es

Используйте следующую команду, чтобы загрузить этот набор данных в TFDS:

ds = tfds.load('huggingface:xtreme/MLQA.es.es')

• Описание :

MLQA (MultiLingual Question Answering) is a benchmark dataset for evaluating cross-lingual question answering performance.
MLQA consists of over 5K extractive QA instances (12K in English) in SQuAD format in seven languages - English, Arabic,
German, Spanish, Hindi, Vietnamese and Simplified Chinese. MLQA is highly parallel, with QA instances parallel between
4 different languages on average.
The Cross-lingual TRansfer Evaluation of Multilingual Encoders (XTREME) benchmark is a benchmark for the evaluation of
the cross-lingual generalization ability of pre-trained multilingual models. It covers 40 typologically diverse languages
(spanning 12 language families) and includes nine tasks that collectively require reasoning about different levels of
syntax and semantics. The languages in XTREME are selected to maximize language diversity, coverage in existing tasks,
and availability of training data. Among these are many under-studied languages, such as the Dravidian languages Tamil
(spoken in southern India, Sri Lanka, and Singapore), Telugu and Malayalam (spoken mainly in southern India), and the
Niger-Congo languages Swahili and Yoruba, spoken in Africa.

• Лицензия : Нет известной лицензии.
• Версия : 1.0.0
• Расколы :

• Функции :

{
    "id": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "title": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "context": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "question": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "answers": {
        "feature": {
            "answer_start": {
                "dtype": "int32",
                "id": null,
                "_type": "Value"
            },
            "text": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            }
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

MLQA.es.hi

Используйте следующую команду, чтобы загрузить этот набор данных в TFDS:

ds = tfds.load('huggingface:xtreme/MLQA.es.hi')

• Описание :

MLQA (MultiLingual Question Answering) is a benchmark dataset for evaluating cross-lingual question answering performance.
MLQA consists of over 5K extractive QA instances (12K in English) in SQuAD format in seven languages - English, Arabic,
German, Spanish, Hindi, Vietnamese and Simplified Chinese. MLQA is highly parallel, with QA instances parallel between
4 different languages on average.
The Cross-lingual TRansfer Evaluation of Multilingual Encoders (XTREME) benchmark is a benchmark for the evaluation of
the cross-lingual generalization ability of pre-trained multilingual models. It covers 40 typologically diverse languages
(spanning 12 language families) and includes nine tasks that collectively require reasoning about different levels of
syntax and semantics. The languages in XTREME are selected to maximize language diversity, coverage in existing tasks,
and availability of training data. Among these are many under-studied languages, such as the Dravidian languages Tamil
(spoken in southern India, Sri Lanka, and Singapore), Telugu and Malayalam (spoken mainly in southern India), and the
Niger-Congo languages Swahili and Yoruba, spoken in Africa.

• Лицензия : Нет известной лицензии.
• Версия : 1.0.0
• Расколы :

• Функции :

{
    "id": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "title": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "context": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "question": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "answers": {
        "feature": {
            "answer_start": {
                "dtype": "int32",
                "id": null,
                "_type": "Value"
            },
            "text": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            }
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

MLQA.hi.ar

Используйте следующую команду, чтобы загрузить этот набор данных в TFDS:

ds = tfds.load('huggingface:xtreme/MLQA.hi.ar')

• Описание :

MLQA (MultiLingual Question Answering) is a benchmark dataset for evaluating cross-lingual question answering performance.
MLQA consists of over 5K extractive QA instances (12K in English) in SQuAD format in seven languages - English, Arabic,
German, Spanish, Hindi, Vietnamese and Simplified Chinese. MLQA is highly parallel, with QA instances parallel between
4 different languages on average.
The Cross-lingual TRansfer Evaluation of Multilingual Encoders (XTREME) benchmark is a benchmark for the evaluation of
the cross-lingual generalization ability of pre-trained multilingual models. It covers 40 typologically diverse languages
(spanning 12 language families) and includes nine tasks that collectively require reasoning about different levels of
syntax and semantics. The languages in XTREME are selected to maximize language diversity, coverage in existing tasks,
and availability of training data. Among these are many under-studied languages, such as the Dravidian languages Tamil
(spoken in southern India, Sri Lanka, and Singapore), Telugu and Malayalam (spoken mainly in southern India), and the
Niger-Congo languages Swahili and Yoruba, spoken in Africa.

• Лицензия : Нет известной лицензии.
• Версия : 1.0.0
• Расколы :

• Функции :

{
    "id": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "title": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "context": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "question": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "answers": {
        "feature": {
            "answer_start": {
                "dtype": "int32",
                "id": null,
                "_type": "Value"
            },
            "text": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            }
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

MLQA.hi.de

Используйте следующую команду, чтобы загрузить этот набор данных в TFDS:

ds = tfds.load('huggingface:xtreme/MLQA.hi.de')

• Описание :

MLQA (MultiLingual Question Answering) is a benchmark dataset for evaluating cross-lingual question answering performance.
MLQA consists of over 5K extractive QA instances (12K in English) in SQuAD format in seven languages - English, Arabic,
German, Spanish, Hindi, Vietnamese and Simplified Chinese. MLQA is highly parallel, with QA instances parallel between
4 different languages on average.
The Cross-lingual TRansfer Evaluation of Multilingual Encoders (XTREME) benchmark is a benchmark for the evaluation of
the cross-lingual generalization ability of pre-trained multilingual models. It covers 40 typologically diverse languages
(spanning 12 language families) and includes nine tasks that collectively require reasoning about different levels of
syntax and semantics. The languages in XTREME are selected to maximize language diversity, coverage in existing tasks,
and availability of training data. Among these are many under-studied languages, such as the Dravidian languages Tamil
(spoken in southern India, Sri Lanka, and Singapore), Telugu and Malayalam (spoken mainly in southern India), and the
Niger-Congo languages Swahili and Yoruba, spoken in Africa.

• Лицензия : Нет известной лицензии.
• Версия : 1.0.0
• Расколы :

• Функции :

{
    "id": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "title": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "context": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "question": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "answers": {
        "feature": {
            "answer_start": {
                "dtype": "int32",
                "id": null,
                "_type": "Value"
            },
            "text": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            }
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

MLQA.hi.vi

Используйте следующую команду, чтобы загрузить этот набор данных в TFDS:

ds = tfds.load('huggingface:xtreme/MLQA.hi.vi')

• Описание :

MLQA (MultiLingual Question Answering) is a benchmark dataset for evaluating cross-lingual question answering performance.
MLQA consists of over 5K extractive QA instances (12K in English) in SQuAD format in seven languages - English, Arabic,
German, Spanish, Hindi, Vietnamese and Simplified Chinese. MLQA is highly parallel, with QA instances parallel between
4 different languages on average.
The Cross-lingual TRansfer Evaluation of Multilingual Encoders (XTREME) benchmark is a benchmark for the evaluation of
the cross-lingual generalization ability of pre-trained multilingual models. It covers 40 typologically diverse languages
(spanning 12 language families) and includes nine tasks that collectively require reasoning about different levels of
syntax and semantics. The languages in XTREME are selected to maximize language diversity, coverage in existing tasks,
and availability of training data. Among these are many under-studied languages, such as the Dravidian languages Tamil
(spoken in southern India, Sri Lanka, and Singapore), Telugu and Malayalam (spoken mainly in southern India), and the
Niger-Congo languages Swahili and Yoruba, spoken in Africa.

• Лицензия : Нет известной лицензии.
• Версия : 1.0.0
• Расколы :

• Функции :

{
    "id": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "title": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "context": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "question": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "answers": {
        "feature": {
            "answer_start": {
                "dtype": "int32",
                "id": null,
                "_type": "Value"
            },
            "text": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            }
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

MLQA.hi.zh

Используйте следующую команду, чтобы загрузить этот набор данных в TFDS:

ds = tfds.load('huggingface:xtreme/MLQA.hi.zh')

• Описание :

MLQA (MultiLingual Question Answering) is a benchmark dataset for evaluating cross-lingual question answering performance.
MLQA consists of over 5K extractive QA instances (12K in English) in SQuAD format in seven languages - English, Arabic,
German, Spanish, Hindi, Vietnamese and Simplified Chinese. MLQA is highly parallel, with QA instances parallel between
4 different languages on average.
The Cross-lingual TRansfer Evaluation of Multilingual Encoders (XTREME) benchmark is a benchmark for the evaluation of
the cross-lingual generalization ability of pre-trained multilingual models. It covers 40 typologically diverse languages
(spanning 12 language families) and includes nine tasks that collectively require reasoning about different levels of
syntax and semantics. The languages in XTREME are selected to maximize language diversity, coverage in existing tasks,
and availability of training data. Among these are many under-studied languages, such as the Dravidian languages Tamil
(spoken in southern India, Sri Lanka, and Singapore), Telugu and Malayalam (spoken mainly in southern India), and the
Niger-Congo languages Swahili and Yoruba, spoken in Africa.

• Лицензия : Нет известной лицензии.
• Версия : 1.0.0
• Расколы :

• Функции :

{
    "id": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "title": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "context": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "question": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "answers": {
        "feature": {
            "answer_start": {
                "dtype": "int32",
                "id": null,
                "_type": "Value"
            },
            "text": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            }
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

MLQA.hi.en

Используйте следующую команду, чтобы загрузить этот набор данных в TFDS:

ds = tfds.load('huggingface:xtreme/MLQA.hi.en')

• Описание :

MLQA (MultiLingual Question Answering) is a benchmark dataset for evaluating cross-lingual question answering performance.
MLQA consists of over 5K extractive QA instances (12K in English) in SQuAD format in seven languages - English, Arabic,
German, Spanish, Hindi, Vietnamese and Simplified Chinese. MLQA is highly parallel, with QA instances parallel between
4 different languages on average.
The Cross-lingual TRansfer Evaluation of Multilingual Encoders (XTREME) benchmark is a benchmark for the evaluation of
the cross-lingual generalization ability of pre-trained multilingual models. It covers 40 typologically diverse languages
(spanning 12 language families) and includes nine tasks that collectively require reasoning about different levels of
syntax and semantics. The languages in XTREME are selected to maximize language diversity, coverage in existing tasks,
and availability of training data. Among these are many under-studied languages, such as the Dravidian languages Tamil
(spoken in southern India, Sri Lanka, and Singapore), Telugu and Malayalam (spoken mainly in southern India), and the
Niger-Congo languages Swahili and Yoruba, spoken in Africa.

• Лицензия : Нет известной лицензии.
• Версия : 1.0.0
• Расколы :

• Функции :

{
    "id": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "title": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "context": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "question": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "answers": {
        "feature": {
            "answer_start": {
                "dtype": "int32",
                "id": null,
                "_type": "Value"
            },
            "text": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            }
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

MLQA.hi.es

Используйте следующую команду, чтобы загрузить этот набор данных в TFDS:

ds = tfds.load('huggingface:xtreme/MLQA.hi.es')

• Описание :

MLQA (MultiLingual Question Answering) is a benchmark dataset for evaluating cross-lingual question answering performance.
MLQA consists of over 5K extractive QA instances (12K in English) in SQuAD format in seven languages - English, Arabic,
German, Spanish, Hindi, Vietnamese and Simplified Chinese. MLQA is highly parallel, with QA instances parallel between
4 different languages on average.
The Cross-lingual TRansfer Evaluation of Multilingual Encoders (XTREME) benchmark is a benchmark for the evaluation of
the cross-lingual generalization ability of pre-trained multilingual models. It covers 40 typologically diverse languages
(spanning 12 language families) and includes nine tasks that collectively require reasoning about different levels of
syntax and semantics. The languages in XTREME are selected to maximize language diversity, coverage in existing tasks,
and availability of training data. Among these are many under-studied languages, such as the Dravidian languages Tamil
(spoken in southern India, Sri Lanka, and Singapore), Telugu and Malayalam (spoken mainly in southern India), and the
Niger-Congo languages Swahili and Yoruba, spoken in Africa.

• Лицензия : Нет известной лицензии.
• Версия : 1.0.0
• Расколы :

• Функции :

{
    "id": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "title": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "context": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "question": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "answers": {
        "feature": {
            "answer_start": {
                "dtype": "int32",
                "id": null,
                "_type": "Value"
            },
            "text": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            }
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

MLQA.hi.hi

Используйте следующую команду, чтобы загрузить этот набор данных в TFDS:

ds = tfds.load('huggingface:xtreme/MLQA.hi.hi')

• Описание :

MLQA (MultiLingual Question Answering) is a benchmark dataset for evaluating cross-lingual question answering performance.
MLQA consists of over 5K extractive QA instances (12K in English) in SQuAD format in seven languages - English, Arabic,
German, Spanish, Hindi, Vietnamese and Simplified Chinese. MLQA is highly parallel, with QA instances parallel between
4 different languages on average.
The Cross-lingual TRansfer Evaluation of Multilingual Encoders (XTREME) benchmark is a benchmark for the evaluation of
the cross-lingual generalization ability of pre-trained multilingual models. It covers 40 typologically diverse languages
(spanning 12 language families) and includes nine tasks that collectively require reasoning about different levels of
syntax and semantics. The languages in XTREME are selected to maximize language diversity, coverage in existing tasks,
and availability of training data. Among these are many under-studied languages, such as the Dravidian languages Tamil
(spoken in southern India, Sri Lanka, and Singapore), Telugu and Malayalam (spoken mainly in southern India), and the
Niger-Congo languages Swahili and Yoruba, spoken in Africa.

• Лицензия : Нет известной лицензии.
• Версия : 1.0.0
• Расколы :

• Функции :

{
    "id": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "title": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "context": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "question": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "answers": {
        "feature": {
            "answer_start": {
                "dtype": "int32",
                "id": null,
                "_type": "Value"
            },
            "text": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            }
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

XQuAD.ar

Используйте следующую команду, чтобы загрузить этот набор данных в TFDS:

ds = tfds.load('huggingface:xtreme/XQuAD.ar')

• Описание :

XQuAD (Cross-lingual Question Answering Dataset) is a benchmark dataset for evaluating cross-lingual question
answering performance. The dataset consists of a subset of 240 paragraphs and 1190 question-answer pairs from
the development set of SQuAD v1.1 (Rajpurkar et al., 2016) together with their professional translations into
ten languages: Spanish, German, Greek, Russian, Turkish, Arabic, Vietnamese, Thai, Chinese, and Hindi. Consequently,
the dataset is entirely parallel across 11 languages.
The Cross-lingual TRansfer Evaluation of Multilingual Encoders (XTREME) benchmark is a benchmark for the evaluation of
the cross-lingual generalization ability of pre-trained multilingual models. It covers 40 typologically diverse languages
(spanning 12 language families) and includes nine tasks that collectively require reasoning about different levels of
syntax and semantics. The languages in XTREME are selected to maximize language diversity, coverage in existing tasks,
and availability of training data. Among these are many under-studied languages, such as the Dravidian languages Tamil
(spoken in southern India, Sri Lanka, and Singapore), Telugu and Malayalam (spoken mainly in southern India), and the
Niger-Congo languages Swahili and Yoruba, spoken in Africa.

• Лицензия : Нет известной лицензии.
• Версия : 1.0.0
• Расколы :

• Функции :

{
    "id": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "context": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "question": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "answers": {
        "feature": {
            "answer_start": {
                "dtype": "int32",
                "id": null,
                "_type": "Value"
            },
            "text": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            }
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

XQuAD.de

Используйте следующую команду, чтобы загрузить этот набор данных в TFDS:

ds = tfds.load('huggingface:xtreme/XQuAD.de')

• Описание :

XQuAD (Cross-lingual Question Answering Dataset) is a benchmark dataset for evaluating cross-lingual question
answering performance. The dataset consists of a subset of 240 paragraphs and 1190 question-answer pairs from
the development set of SQuAD v1.1 (Rajpurkar et al., 2016) together with their professional translations into
ten languages: Spanish, German, Greek, Russian, Turkish, Arabic, Vietnamese, Thai, Chinese, and Hindi. Consequently,
the dataset is entirely parallel across 11 languages.
The Cross-lingual TRansfer Evaluation of Multilingual Encoders (XTREME) benchmark is a benchmark for the evaluation of
the cross-lingual generalization ability of pre-trained multilingual models. It covers 40 typologically diverse languages
(spanning 12 language families) and includes nine tasks that collectively require reasoning about different levels of
syntax and semantics. The languages in XTREME are selected to maximize language diversity, coverage in existing tasks,
and availability of training data. Among these are many under-studied languages, such as the Dravidian languages Tamil
(spoken in southern India, Sri Lanka, and Singapore), Telugu and Malayalam (spoken mainly in southern India), and the
Niger-Congo languages Swahili and Yoruba, spoken in Africa.

• Лицензия : Нет известной лицензии.
• Версия : 1.0.0
• Расколы :

• Функции :

{
    "id": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "context": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "question": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "answers": {
        "feature": {
            "answer_start": {
                "dtype": "int32",
                "id": null,
                "_type": "Value"
            },
            "text": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            }
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

XQuAD.vi

Используйте следующую команду, чтобы загрузить этот набор данных в TFDS:

ds = tfds.load('huggingface:xtreme/XQuAD.vi')

• Описание :

XQuAD (Cross-lingual Question Answering Dataset) is a benchmark dataset for evaluating cross-lingual question
answering performance. The dataset consists of a subset of 240 paragraphs and 1190 question-answer pairs from
the development set of SQuAD v1.1 (Rajpurkar et al., 2016) together with their professional translations into
ten languages: Spanish, German, Greek, Russian, Turkish, Arabic, Vietnamese, Thai, Chinese, and Hindi. Consequently,
the dataset is entirely parallel across 11 languages.
The Cross-lingual TRansfer Evaluation of Multilingual Encoders (XTREME) benchmark is a benchmark for the evaluation of
the cross-lingual generalization ability of pre-trained multilingual models. It covers 40 typologically diverse languages
(spanning 12 language families) and includes nine tasks that collectively require reasoning about different levels of
syntax and semantics. The languages in XTREME are selected to maximize language diversity, coverage in existing tasks,
and availability of training data. Among these are many under-studied languages, such as the Dravidian languages Tamil
(spoken in southern India, Sri Lanka, and Singapore), Telugu and Malayalam (spoken mainly in southern India), and the
Niger-Congo languages Swahili and Yoruba, spoken in Africa.

• Лицензия : Нет известной лицензии.
• Версия : 1.0.0
• Расколы :

• Функции :

{
    "id": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "context": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "question": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "answers": {
        "feature": {
            "answer_start": {
                "dtype": "int32",
                "id": null,
                "_type": "Value"
            },
            "text": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            }
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

XQuAD.ж

Используйте следующую команду, чтобы загрузить этот набор данных в TFDS:

ds = tfds.load('huggingface:xtreme/XQuAD.zh')

• Описание :

XQuAD (Cross-lingual Question Answering Dataset) is a benchmark dataset for evaluating cross-lingual question
answering performance. The dataset consists of a subset of 240 paragraphs and 1190 question-answer pairs from
the development set of SQuAD v1.1 (Rajpurkar et al., 2016) together with their professional translations into
ten languages: Spanish, German, Greek, Russian, Turkish, Arabic, Vietnamese, Thai, Chinese, and Hindi. Consequently,
the dataset is entirely parallel across 11 languages.
The Cross-lingual TRansfer Evaluation of Multilingual Encoders (XTREME) benchmark is a benchmark for the evaluation of
the cross-lingual generalization ability of pre-trained multilingual models. It covers 40 typologically diverse languages
(spanning 12 language families) and includes nine tasks that collectively require reasoning about different levels of
syntax and semantics. The languages in XTREME are selected to maximize language diversity, coverage in existing tasks,
and availability of training data. Among these are many under-studied languages, such as the Dravidian languages Tamil
(spoken in southern India, Sri Lanka, and Singapore), Telugu and Malayalam (spoken mainly in southern India), and the
Niger-Congo languages Swahili and Yoruba, spoken in Africa.

• Лицензия : Нет известной лицензии.
• Версия : 1.0.0
• Расколы :

• Функции :

{
    "id": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "context": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "question": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "answers": {
        "feature": {
            "answer_start": {
                "dtype": "int32",
                "id": null,
                "_type": "Value"
            },
            "text": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            }
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

XQuAD.ru

Используйте следующую команду, чтобы загрузить этот набор данных в TFDS:

ds = tfds.load('huggingface:xtreme/XQuAD.en')

• Описание :

XQuAD (Cross-lingual Question Answering Dataset) is a benchmark dataset for evaluating cross-lingual question
answering performance. The dataset consists of a subset of 240 paragraphs and 1190 question-answer pairs from
the development set of SQuAD v1.1 (Rajpurkar et al., 2016) together with their professional translations into
ten languages: Spanish, German, Greek, Russian, Turkish, Arabic, Vietnamese, Thai, Chinese, and Hindi. Consequently,
the dataset is entirely parallel across 11 languages.
The Cross-lingual TRansfer Evaluation of Multilingual Encoders (XTREME) benchmark is a benchmark for the evaluation of
the cross-lingual generalization ability of pre-trained multilingual models. It covers 40 typologically diverse languages
(spanning 12 language families) and includes nine tasks that collectively require reasoning about different levels of
syntax and semantics. The languages in XTREME are selected to maximize language diversity, coverage in existing tasks,
and availability of training data. Among these are many under-studied languages, such as the Dravidian languages Tamil
(spoken in southern India, Sri Lanka, and Singapore), Telugu and Malayalam (spoken mainly in southern India), and the
Niger-Congo languages Swahili and Yoruba, spoken in Africa.

• Лицензия : Нет известной лицензии.
• Версия : 1.0.0
• Расколы :

• Функции :

{
    "id": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "context": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "question": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "answers": {
        "feature": {
            "answer_start": {
                "dtype": "int32",
                "id": null,
                "_type": "Value"
            },
            "text": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            }
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

XQuAD.es

Используйте следующую команду, чтобы загрузить этот набор данных в TFDS:

ds = tfds.load('huggingface:xtreme/XQuAD.es')

• Описание :

XQuAD (Cross-lingual Question Answering Dataset) is a benchmark dataset for evaluating cross-lingual question
answering performance. The dataset consists of a subset of 240 paragraphs and 1190 question-answer pairs from
the development set of SQuAD v1.1 (Rajpurkar et al., 2016) together with their professional translations into
ten languages: Spanish, German, Greek, Russian, Turkish, Arabic, Vietnamese, Thai, Chinese, and Hindi. Consequently,
the dataset is entirely parallel across 11 languages.
The Cross-lingual TRansfer Evaluation of Multilingual Encoders (XTREME) benchmark is a benchmark for the evaluation of
the cross-lingual generalization ability of pre-trained multilingual models. It covers 40 typologically diverse languages
(spanning 12 language families) and includes nine tasks that collectively require reasoning about different levels of
syntax and semantics. The languages in XTREME are selected to maximize language diversity, coverage in existing tasks,
and availability of training data. Among these are many under-studied languages, such as the Dravidian languages Tamil
(spoken in southern India, Sri Lanka, and Singapore), Telugu and Malayalam (spoken mainly in southern India), and the
Niger-Congo languages Swahili and Yoruba, spoken in Africa.

• Лицензия : Нет известной лицензии.
• Версия : 1.0.0
• Расколы :

• Функции :

{
    "id": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "context": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "question": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "answers": {
        "feature": {
            "answer_start": {
                "dtype": "int32",
                "id": null,
                "_type": "Value"
            },
            "text": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            }
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

XQuAD.hi

Используйте следующую команду, чтобы загрузить этот набор данных в TFDS:

ds = tfds.load('huggingface:xtreme/XQuAD.hi')

• Описание :

XQuAD (Cross-lingual Question Answering Dataset) is a benchmark dataset for evaluating cross-lingual question
answering performance. The dataset consists of a subset of 240 paragraphs and 1190 question-answer pairs from
the development set of SQuAD v1.1 (Rajpurkar et al., 2016) together with their professional translations into
ten languages: Spanish, German, Greek, Russian, Turkish, Arabic, Vietnamese, Thai, Chinese, and Hindi. Consequently,
the dataset is entirely parallel across 11 languages.
The Cross-lingual TRansfer Evaluation of Multilingual Encoders (XTREME) benchmark is a benchmark for the evaluation of
the cross-lingual generalization ability of pre-trained multilingual models. It covers 40 typologically diverse languages
(spanning 12 language families) and includes nine tasks that collectively require reasoning about different levels of
syntax and semantics. The languages in XTREME are selected to maximize language diversity, coverage in existing tasks,
and availability of training data. Among these are many under-studied languages, such as the Dravidian languages Tamil
(spoken in southern India, Sri Lanka, and Singapore), Telugu and Malayalam (spoken mainly in southern India), and the
Niger-Congo languages Swahili and Yoruba, spoken in Africa.

• Лицензия : Нет известной лицензии.
• Версия : 1.0.0
• Расколы :

• Функции :

{
    "id": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "context": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "question": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "answers": {
        "feature": {
            "answer_start": {
                "dtype": "int32",
                "id": null,
                "_type": "Value"
            },
            "text": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            }
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

XQuAD.el

Используйте следующую команду, чтобы загрузить этот набор данных в TFDS:

ds = tfds.load('huggingface:xtreme/XQuAD.el')

• Описание :

XQuAD (Cross-lingual Question Answering Dataset) is a benchmark dataset for evaluating cross-lingual question
answering performance. The dataset consists of a subset of 240 paragraphs and 1190 question-answer pairs from
the development set of SQuAD v1.1 (Rajpurkar et al., 2016) together with their professional translations into
ten languages: Spanish, German, Greek, Russian, Turkish, Arabic, Vietnamese, Thai, Chinese, and Hindi. Consequently,
the dataset is entirely parallel across 11 languages.
The Cross-lingual TRansfer Evaluation of Multilingual Encoders (XTREME) benchmark is a benchmark for the evaluation of
the cross-lingual generalization ability of pre-trained multilingual models. It covers 40 typologically diverse languages
(spanning 12 language families) and includes nine tasks that collectively require reasoning about different levels of
syntax and semantics. The languages in XTREME are selected to maximize language diversity, coverage in existing tasks,
and availability of training data. Among these are many under-studied languages, such as the Dravidian languages Tamil
(spoken in southern India, Sri Lanka, and Singapore), Telugu and Malayalam (spoken mainly in southern India), and the
Niger-Congo languages Swahili and Yoruba, spoken in Africa.

• Лицензия : Нет известной лицензии.
• Версия : 1.0.0
• Расколы :

• Функции :

{
    "id": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "context": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "question": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "answers": {
        "feature": {
            "answer_start": {
                "dtype": "int32",
                "id": null,
                "_type": "Value"
            },
            "text": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            }
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

XQuAD.ru

Используйте следующую команду, чтобы загрузить этот набор данных в TFDS:

ds = tfds.load('huggingface:xtreme/XQuAD.ru')

• Описание :

XQuAD (Cross-lingual Question Answering Dataset) is a benchmark dataset for evaluating cross-lingual question
answering performance. The dataset consists of a subset of 240 paragraphs and 1190 question-answer pairs from
the development set of SQuAD v1.1 (Rajpurkar et al., 2016) together with their professional translations into
ten languages: Spanish, German, Greek, Russian, Turkish, Arabic, Vietnamese, Thai, Chinese, and Hindi. Consequently,
the dataset is entirely parallel across 11 languages.
The Cross-lingual TRansfer Evaluation of Multilingual Encoders (XTREME) benchmark is a benchmark for the evaluation of
the cross-lingual generalization ability of pre-trained multilingual models. It covers 40 typologically diverse languages
(spanning 12 language families) and includes nine tasks that collectively require reasoning about different levels of
syntax and semantics. The languages in XTREME are selected to maximize language diversity, coverage in existing tasks,
and availability of training data. Among these are many under-studied languages, such as the Dravidian languages Tamil
(spoken in southern India, Sri Lanka, and Singapore), Telugu and Malayalam (spoken mainly in southern India), and the
Niger-Congo languages Swahili and Yoruba, spoken in Africa.

• Лицензия : Нет известной лицензии.
• Версия : 1.0.0
• Расколы :

• Функции :

{
    "id": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "context": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "question": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "answers": {
        "feature": {
            "answer_start": {
                "dtype": "int32",
                "id": null,
                "_type": "Value"
            },
            "text": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            }
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

XQuAD.th

Используйте следующую команду, чтобы загрузить этот набор данных в TFDS:

ds = tfds.load('huggingface:xtreme/XQuAD.th')

• Описание :

XQuAD (Cross-lingual Question Answering Dataset) is a benchmark dataset for evaluating cross-lingual question
answering performance. The dataset consists of a subset of 240 paragraphs and 1190 question-answer pairs from
the development set of SQuAD v1.1 (Rajpurkar et al., 2016) together with their professional translations into
ten languages: Spanish, German, Greek, Russian, Turkish, Arabic, Vietnamese, Thai, Chinese, and Hindi. Consequently,
the dataset is entirely parallel across 11 languages.
The Cross-lingual TRansfer Evaluation of Multilingual Encoders (XTREME) benchmark is a benchmark for the evaluation of
the cross-lingual generalization ability of pre-trained multilingual models. It covers 40 typologically diverse languages
(spanning 12 language families) and includes nine tasks that collectively require reasoning about different levels of
syntax and semantics. The languages in XTREME are selected to maximize language diversity, coverage in existing tasks,
and availability of training data. Among these are many under-studied languages, such as the Dravidian languages Tamil
(spoken in southern India, Sri Lanka, and Singapore), Telugu and Malayalam (spoken mainly in southern India), and the
Niger-Congo languages Swahili and Yoruba, spoken in Africa.

• Лицензия : Нет известной лицензии.
• Версия : 1.0.0
• Расколы :

• Функции :

{
    "id": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "context": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "question": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "answers": {
        "feature": {
            "answer_start": {
                "dtype": "int32",
                "id": null,
                "_type": "Value"
            },
            "text": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            }
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

XQuAD.tr

Используйте следующую команду, чтобы загрузить этот набор данных в TFDS:

ds = tfds.load('huggingface:xtreme/XQuAD.tr')

• Описание :

XQuAD (Cross-lingual Question Answering Dataset) is a benchmark dataset for evaluating cross-lingual question
answering performance. The dataset consists of a subset of 240 paragraphs and 1190 question-answer pairs from
the development set of SQuAD v1.1 (Rajpurkar et al., 2016) together with their professional translations into
ten languages: Spanish, German, Greek, Russian, Turkish, Arabic, Vietnamese, Thai, Chinese, and Hindi. Consequently,
the dataset is entirely parallel across 11 languages.
The Cross-lingual TRansfer Evaluation of Multilingual Encoders (XTREME) benchmark is a benchmark for the evaluation of
the cross-lingual generalization ability of pre-trained multilingual models. It covers 40 typologically diverse languages
(spanning 12 language families) and includes nine tasks that collectively require reasoning about different levels of
syntax and semantics. The languages in XTREME are selected to maximize language diversity, coverage in existing tasks,
and availability of training data. Among these are many under-studied languages, such as the Dravidian languages Tamil
(spoken in southern India, Sri Lanka, and Singapore), Telugu and Malayalam (spoken mainly in southern India), and the
Niger-Congo languages Swahili and Yoruba, spoken in Africa.

• Лицензия : Нет известной лицензии.
• Версия : 1.0.0
• Расколы :

• Функции :

{
    "id": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "context": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "question": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "answers": {
        "feature": {
            "answer_start": {
                "dtype": "int32",
                "id": null,
                "_type": "Value"
            },
            "text": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            }
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

bucc18.de

Используйте следующую команду, чтобы загрузить этот набор данных в TFDS:

ds = tfds.load('huggingface:xtreme/bucc18.de')

• Описание :

Building and Using Comparable Corpora

The Cross-lingual TRansfer Evaluation of Multilingual Encoders (XTREME) benchmark is a benchmark for the evaluation of
the cross-lingual generalization ability of pre-trained multilingual models. It covers 40 typologically diverse languages
(spanning 12 language families) and includes nine tasks that collectively require reasoning about different levels of
syntax and semantics. The languages in XTREME are selected to maximize language diversity, coverage in existing tasks,
and availability of training data. Among these are many under-studied languages, such as the Dravidian languages Tamil
(spoken in southern India, Sri Lanka, and Singapore), Telugu and Malayalam (spoken mainly in southern India), and the
Niger-Congo languages Swahili and Yoruba, spoken in Africa.

• Лицензия : Нет известной лицензии.
• Версия : 1.0.0
• Расколы :

• Функции :

{
    "source_sentence": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "target_sentence": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "source_lang": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "target_lang": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

bucc18.fr

Используйте следующую команду, чтобы загрузить этот набор данных в TFDS:

ds = tfds.load('huggingface:xtreme/bucc18.fr')

• Описание :

Building and Using Comparable Corpora

The Cross-lingual TRansfer Evaluation of Multilingual Encoders (XTREME) benchmark is a benchmark for the evaluation of
the cross-lingual generalization ability of pre-trained multilingual models. It covers 40 typologically diverse languages
(spanning 12 language families) and includes nine tasks that collectively require reasoning about different levels of
syntax and semantics. The languages in XTREME are selected to maximize language diversity, coverage in existing tasks,
and availability of training data. Among these are many under-studied languages, such as the Dravidian languages Tamil
(spoken in southern India, Sri Lanka, and Singapore), Telugu and Malayalam (spoken mainly in southern India), and the
Niger-Congo languages Swahili and Yoruba, spoken in Africa.

• Лицензия : Нет известной лицензии.
• Версия : 1.0.0
• Расколы :

• Функции :

{
    "source_sentence": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "target_sentence": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "source_lang": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "target_lang": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

bucc18.zh

Используйте следующую команду, чтобы загрузить этот набор данных в TFDS:

ds = tfds.load('huggingface:xtreme/bucc18.zh')

• Описание :

Building and Using Comparable Corpora

The Cross-lingual TRansfer Evaluation of Multilingual Encoders (XTREME) benchmark is a benchmark for the evaluation of
the cross-lingual generalization ability of pre-trained multilingual models. It covers 40 typologically diverse languages
(spanning 12 language families) and includes nine tasks that collectively require reasoning about different levels of
syntax and semantics. The languages in XTREME are selected to maximize language diversity, coverage in existing tasks,
and availability of training data. Among these are many under-studied languages, such as the Dravidian languages Tamil
(spoken in southern India, Sri Lanka, and Singapore), Telugu and Malayalam (spoken mainly in southern India), and the
Niger-Congo languages Swahili and Yoruba, spoken in Africa.

• Лицензия : Нет известной лицензии.
• Версия : 1.0.0
• Расколы :

• Функции :

{
    "source_sentence": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "target_sentence": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "source_lang": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "target_lang": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

bucc18.ru

Используйте следующую команду, чтобы загрузить этот набор данных в TFDS:

ds = tfds.load('huggingface:xtreme/bucc18.ru')

• Описание :

Building and Using Comparable Corpora

The Cross-lingual TRansfer Evaluation of Multilingual Encoders (XTREME) benchmark is a benchmark for the evaluation of
the cross-lingual generalization ability of pre-trained multilingual models. It covers 40 typologically diverse languages
(spanning 12 language families) and includes nine tasks that collectively require reasoning about different levels of
syntax and semantics. The languages in XTREME are selected to maximize language diversity, coverage in existing tasks,
and availability of training data. Among these are many under-studied languages, such as the Dravidian languages Tamil
(spoken in southern India, Sri Lanka, and Singapore), Telugu and Malayalam (spoken mainly in southern India), and the
Niger-Congo languages Swahili and Yoruba, spoken in Africa.

• Лицензия : Нет известной лицензии.
• Версия : 1.0.0
• Расколы :

• Функции :

{
    "source_sentence": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "target_sentence": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "source_lang": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "target_lang": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

PAWS-X.de

Используйте следующую команду, чтобы загрузить этот набор данных в TFDS:

ds = tfds.load('huggingface:xtreme/PAWS-X.de')

• Описание :

This dataset contains 23,659 human translated PAWS evaluation pairs and 296,406 machine translated training
pairs in six typologically distinct languages: French, Spanish, German, Chinese, Japanese, and Korean. All
translated pairs are sourced from examples in PAWS-Wiki.
The Cross-lingual TRansfer Evaluation of Multilingual Encoders (XTREME) benchmark is a benchmark for the evaluation of
the cross-lingual generalization ability of pre-trained multilingual models. It covers 40 typologically diverse languages
(spanning 12 language families) and includes nine tasks that collectively require reasoning about different levels of
syntax and semantics. The languages in XTREME are selected to maximize language diversity, coverage in existing tasks,
and availability of training data. Among these are many under-studied languages, such as the Dravidian languages Tamil
(spoken in southern India, Sri Lanka, and Singapore), Telugu and Malayalam (spoken mainly in southern India), and the
Niger-Congo languages Swahili and Yoruba, spoken in Africa.

• Лицензия : Нет известной лицензии.
• Версия : 1.0.0
• Расколы :

• Функции :

{
    "sentence1": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "sentence2": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "label": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

PAWS-X.ru

Используйте следующую команду, чтобы загрузить этот набор данных в TFDS:

ds = tfds.load('huggingface:xtreme/PAWS-X.en')

• Описание :

This dataset contains 23,659 human translated PAWS evaluation pairs and 296,406 machine translated training
pairs in six typologically distinct languages: French, Spanish, German, Chinese, Japanese, and Korean. All
translated pairs are sourced from examples in PAWS-Wiki.
The Cross-lingual TRansfer Evaluation of Multilingual Encoders (XTREME) benchmark is a benchmark for the evaluation of
the cross-lingual generalization ability of pre-trained multilingual models. It covers 40 typologically diverse languages
(spanning 12 language families) and includes nine tasks that collectively require reasoning about different levels of
syntax and semantics. The languages in XTREME are selected to maximize language diversity, coverage in existing tasks,
and availability of training data. Among these are many under-studied languages, such as the Dravidian languages Tamil
(spoken in southern India, Sri Lanka, and Singapore), Telugu and Malayalam (spoken mainly in southern India), and the
Niger-Congo languages Swahili and Yoruba, spoken in Africa.

• Лицензия : Нет известной лицензии.
• Версия : 1.0.0
• Расколы :

• Функции :

{
    "sentence1": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "sentence2": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "label": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

PAWS-X.es

Используйте следующую команду, чтобы загрузить этот набор данных в TFDS:

ds = tfds.load('huggingface:xtreme/PAWS-X.es')

• Описание :

This dataset contains 23,659 human translated PAWS evaluation pairs and 296,406 machine translated training
pairs in six typologically distinct languages: French, Spanish, German, Chinese, Japanese, and Korean. All
translated pairs are sourced from examples in PAWS-Wiki.
The Cross-lingual TRansfer Evaluation of Multilingual Encoders (XTREME) benchmark is a benchmark for the evaluation of
the cross-lingual generalization ability of pre-trained multilingual models. It covers 40 typologically diverse languages
(spanning 12 language families) and includes nine tasks that collectively require reasoning about different levels of
syntax and semantics. The languages in XTREME are selected to maximize language diversity, coverage in existing tasks,
and availability of training data. Among these are many under-studied languages, such as the Dravidian languages Tamil
(spoken in southern India, Sri Lanka, and Singapore), Telugu and Malayalam (spoken mainly in southern India), and the
Niger-Congo languages Swahili and Yoruba, spoken in Africa.

• Лицензия : Нет известной лицензии.
• Версия : 1.0.0
• Расколы :

• Функции :

{
    "sentence1": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "sentence2": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "label": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

ЛАПЫ-X.fr

Используйте следующую команду, чтобы загрузить этот набор данных в TFDS:

ds = tfds.load('huggingface:xtreme/PAWS-X.fr')

• Описание :

This dataset contains 23,659 human translated PAWS evaluation pairs and 296,406 machine translated training
pairs in six typologically distinct languages: French, Spanish, German, Chinese, Japanese, and Korean. All
translated pairs are sourced from examples in PAWS-Wiki.
The Cross-lingual TRansfer Evaluation of Multilingual Encoders (XTREME) benchmark is a benchmark for the evaluation of
the cross-lingual generalization ability of pre-trained multilingual models. It covers 40 typologically diverse languages
(spanning 12 language families) and includes nine tasks that collectively require reasoning about different levels of
syntax and semantics. The languages in XTREME are selected to maximize language diversity, coverage in existing tasks,
and availability of training data. Among these are many under-studied languages, such as the Dravidian languages Tamil
(spoken in southern India, Sri Lanka, and Singapore), Telugu and Malayalam (spoken mainly in southern India), and the
Niger-Congo languages Swahili and Yoruba, spoken in Africa.

• Лицензия : Нет известной лицензии.
• Версия : 1.0.0
• Расколы :

• Функции :

{
    "sentence1": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "sentence2": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "label": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

ЛАПЫ-X.ja

Используйте следующую команду, чтобы загрузить этот набор данных в TFDS:

ds = tfds.load('huggingface:xtreme/PAWS-X.ja')

• Описание :

This dataset contains 23,659 human translated PAWS evaluation pairs and 296,406 machine translated training
pairs in six typologically distinct languages: French, Spanish, German, Chinese, Japanese, and Korean. All
translated pairs are sourced from examples in PAWS-Wiki.
The Cross-lingual TRansfer Evaluation of Multilingual Encoders (XTREME) benchmark is a benchmark for the evaluation of
the cross-lingual generalization ability of pre-trained multilingual models. It covers 40 typologically diverse languages
(spanning 12 language families) and includes nine tasks that collectively require reasoning about different levels of
syntax and semantics. The languages in XTREME are selected to maximize language diversity, coverage in existing tasks,
and availability of training data. Among these are many under-studied languages, such as the Dravidian languages Tamil
(spoken in southern India, Sri Lanka, and Singapore), Telugu and Malayalam (spoken mainly in southern India), and the
Niger-Congo languages Swahili and Yoruba, spoken in Africa.

• Лицензия : Нет известной лицензии.
• Версия : 1.0.0
• Расколы :

• Функции :

{
    "sentence1": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "sentence2": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "label": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

ЛАПЫ-X.ko

Используйте следующую команду, чтобы загрузить этот набор данных в TFDS:

ds = tfds.load('huggingface:xtreme/PAWS-X.ko')

• Описание :

This dataset contains 23,659 human translated PAWS evaluation pairs and 296,406 machine translated training
pairs in six typologically distinct languages: French, Spanish, German, Chinese, Japanese, and Korean. All
translated pairs are sourced from examples in PAWS-Wiki.
The Cross-lingual TRansfer Evaluation of Multilingual Encoders (XTREME) benchmark is a benchmark for the evaluation of
the cross-lingual generalization ability of pre-trained multilingual models. It covers 40 typologically diverse languages
(spanning 12 language families) and includes nine tasks that collectively require reasoning about different levels of
syntax and semantics. The languages in XTREME are selected to maximize language diversity, coverage in existing tasks,
and availability of training data. Among these are many under-studied languages, such as the Dravidian languages Tamil
(spoken in southern India, Sri Lanka, and Singapore), Telugu and Malayalam (spoken mainly in southern India), and the
Niger-Congo languages Swahili and Yoruba, spoken in Africa.

• Лицензия : Нет известной лицензии.
• Версия : 1.0.0
• Расколы :

• Функции :

{
    "sentence1": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "sentence2": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "label": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

ЛАПЫ-X.ж

Используйте следующую команду, чтобы загрузить этот набор данных в TFDS:

ds = tfds.load('huggingface:xtreme/PAWS-X.zh')

• Описание :

This dataset contains 23,659 human translated PAWS evaluation pairs and 296,406 machine translated training
pairs in six typologically distinct languages: French, Spanish, German, Chinese, Japanese, and Korean. All
translated pairs are sourced from examples in PAWS-Wiki.
The Cross-lingual TRansfer Evaluation of Multilingual Encoders (XTREME) benchmark is a benchmark for the evaluation of
the cross-lingual generalization ability of pre-trained multilingual models. It covers 40 typologically diverse languages
(spanning 12 language families) and includes nine tasks that collectively require reasoning about different levels of
syntax and semantics. The languages in XTREME are selected to maximize language diversity, coverage in existing tasks,
and availability of training data. Among these are many under-studied languages, such as the Dravidian languages Tamil
(spoken in southern India, Sri Lanka, and Singapore), Telugu and Malayalam (spoken mainly in southern India), and the
Niger-Congo languages Swahili and Yoruba, spoken in Africa.

• Лицензия : Нет известной лицензии.
• Версия : 1.0.0
• Расколы :

• Функции :

{
    "sentence1": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "sentence2": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "label": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

tatoeba.afr

Используйте следующую команду, чтобы загрузить этот набор данных в TFDS:

ds = tfds.load('huggingface:xtreme/tatoeba.afr')

• Описание :

his data is extracted from the Tatoeba corpus, dated Saturday 2018/11/17.
For each languages, we have selected 1000 English sentences and their translations, if available. Please check
this paper for a description of the languages, their families and scripts as well as baseline results.
Please note that the English sentences are not identical for all language pairs. This means that the results are
not directly comparable across languages. In particular, the sentences tend to have less variety for several
low-resource languages, e.g. "Tom needed water", "Tom needs water", "Tom is getting water", ...

The Cross-lingual TRansfer Evaluation of Multilingual Encoders (XTREME) benchmark is a benchmark for the evaluation of
the cross-lingual generalization ability of pre-trained multilingual models. It covers 40 typologically diverse languages
(spanning 12 language families) and includes nine tasks that collectively require reasoning about different levels of
syntax and semantics. The languages in XTREME are selected to maximize language diversity, coverage in existing tasks,
and availability of training data. Among these are many under-studied languages, such as the Dravidian languages Tamil
(spoken in southern India, Sri Lanka, and Singapore), Telugu and Malayalam (spoken mainly in southern India), and the
Niger-Congo languages Swahili and Yoruba, spoken in Africa.

• Лицензия : Нет известной лицензии.
• Версия : 1.0.0
• Расколы :

• Функции :

{
    "source_sentence": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "target_sentence": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "source_lang": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "target_lang": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

tatoeba.ara

Используйте следующую команду, чтобы загрузить этот набор данных в TFDS:

ds = tfds.load('huggingface:xtreme/tatoeba.ara')

• Описание :

his data is extracted from the Tatoeba corpus, dated Saturday 2018/11/17.
For each languages, we have selected 1000 English sentences and their translations, if available. Please check
this paper for a description of the languages, their families and scripts as well as baseline results.
Please note that the English sentences are not identical for all language pairs. This means that the results are
not directly comparable across languages. In particular, the sentences tend to have less variety for several
low-resource languages, e.g. "Tom needed water", "Tom needs water", "Tom is getting water", ...

The Cross-lingual TRansfer Evaluation of Multilingual Encoders (XTREME) benchmark is a benchmark for the evaluation of
the cross-lingual generalization ability of pre-trained multilingual models. It covers 40 typologically diverse languages
(spanning 12 language families) and includes nine tasks that collectively require reasoning about different levels of
syntax and semantics. The languages in XTREME are selected to maximize language diversity, coverage in existing tasks,
and availability of training data. Among these are many under-studied languages, such as the Dravidian languages Tamil
(spoken in southern India, Sri Lanka, and Singapore), Telugu and Malayalam (spoken mainly in southern India), and the
Niger-Congo languages Swahili and Yoruba, spoken in Africa.

• Лицензия : Нет известной лицензии.
• Версия : 1.0.0
• Расколы :

• Функции :

{
    "source_sentence": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "target_sentence": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "source_lang": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "target_lang": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

tatoeba.ben

Используйте следующую команду, чтобы загрузить этот набор данных в TFDS:

ds = tfds.load('huggingface:xtreme/tatoeba.ben')

• Описание :

his data is extracted from the Tatoeba corpus, dated Saturday 2018/11/17.
For each languages, we have selected 1000 English sentences and their translations, if available. Please check
this paper for a description of the languages, their families and scripts as well as baseline results.
Please note that the English sentences are not identical for all language pairs. This means that the results are
not directly comparable across languages. In particular, the sentences tend to have less variety for several
low-resource languages, e.g. "Tom needed water", "Tom needs water", "Tom is getting water", ...

The Cross-lingual TRansfer Evaluation of Multilingual Encoders (XTREME) benchmark is a benchmark for the evaluation of
the cross-lingual generalization ability of pre-trained multilingual models. It covers 40 typologically diverse languages
(spanning 12 language families) and includes nine tasks that collectively require reasoning about different levels of
syntax and semantics. The languages in XTREME are selected to maximize language diversity, coverage in existing tasks,
and availability of training data. Among these are many under-studied languages, such as the Dravidian languages Tamil
(spoken in southern India, Sri Lanka, and Singapore), Telugu and Malayalam (spoken mainly in southern India), and the
Niger-Congo languages Swahili and Yoruba, spoken in Africa.

• Лицензия : Нет известной лицензии.
• Версия : 1.0.0
• Расколы :

• Функции :

{
    "source_sentence": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "target_sentence": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "source_lang": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "target_lang": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

tatoeba.bul

Используйте следующую команду, чтобы загрузить этот набор данных в TFDS:

ds = tfds.load('huggingface:xtreme/tatoeba.bul')

• Описание :

his data is extracted from the Tatoeba corpus, dated Saturday 2018/11/17.
For each languages, we have selected 1000 English sentences and their translations, if available. Please check
this paper for a description of the languages, their families and scripts as well as baseline results.
Please note that the English sentences are not identical for all language pairs. This means that the results are
not directly comparable across languages. In particular, the sentences tend to have less variety for several
low-resource languages, e.g. "Tom needed water", "Tom needs water", "Tom is getting water", ...

The Cross-lingual TRansfer Evaluation of Multilingual Encoders (XTREME) benchmark is a benchmark for the evaluation of
the cross-lingual generalization ability of pre-trained multilingual models. It covers 40 typologically diverse languages
(spanning 12 language families) and includes nine tasks that collectively require reasoning about different levels of
syntax and semantics. The languages in XTREME are selected to maximize language diversity, coverage in existing tasks,
and availability of training data. Among these are many under-studied languages, such as the Dravidian languages Tamil
(spoken in southern India, Sri Lanka, and Singapore), Telugu and Malayalam (spoken mainly in southern India), and the
Niger-Congo languages Swahili and Yoruba, spoken in Africa.

• Лицензия : Нет известной лицензии.
• Версия : 1.0.0
• Расколы :

• Функции :

{
    "source_sentence": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "target_sentence": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "source_lang": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "target_lang": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

tatoeba.deu

Используйте следующую команду, чтобы загрузить этот набор данных в TFDS:

ds = tfds.load('huggingface:xtreme/tatoeba.deu')

• Описание :

his data is extracted from the Tatoeba corpus, dated Saturday 2018/11/17.
For each languages, we have selected 1000 English sentences and their translations, if available. Please check
this paper for a description of the languages, their families and scripts as well as baseline results.
Please note that the English sentences are not identical for all language pairs. This means that the results are
not directly comparable across languages. In particular, the sentences tend to have less variety for several
low-resource languages, e.g. "Tom needed water", "Tom needs water", "Tom is getting water", ...

The Cross-lingual TRansfer Evaluation of Multilingual Encoders (XTREME) benchmark is a benchmark for the evaluation of
the cross-lingual generalization ability of pre-trained multilingual models. It covers 40 typologically diverse languages
(spanning 12 language families) and includes nine tasks that collectively require reasoning about different levels of
syntax and semantics. The languages in XTREME are selected to maximize language diversity, coverage in existing tasks,
and availability of training data. Among these are many under-studied languages, such as the Dravidian languages Tamil
(spoken in southern India, Sri Lanka, and Singapore), Telugu and Malayalam (spoken mainly in southern India), and the
Niger-Congo languages Swahili and Yoruba, spoken in Africa.

• Лицензия : Нет известной лицензии.
• Версия : 1.0.0
• Расколы :

• Функции :

{
    "source_sentence": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "target_sentence": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "source_lang": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "target_lang": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

tatoeba.cmn

Используйте следующую команду, чтобы загрузить этот набор данных в TFDS:

ds = tfds.load('huggingface:xtreme/tatoeba.cmn')

• Описание :

his data is extracted from the Tatoeba corpus, dated Saturday 2018/11/17.
For each languages, we have selected 1000 English sentences and their translations, if available. Please check
this paper for a description of the languages, their families and scripts as well as baseline results.
Please note that the English sentences are not identical for all language pairs. This means that the results are
not directly comparable across languages. In particular, the sentences tend to have less variety for several
low-resource languages, e.g. "Tom needed water", "Tom needs water", "Tom is getting water", ...

The Cross-lingual TRansfer Evaluation of Multilingual Encoders (XTREME) benchmark is a benchmark for the evaluation of
the cross-lingual generalization ability of pre-trained multilingual models. It covers 40 typologically diverse languages
(spanning 12 language families) and includes nine tasks that collectively require reasoning about different levels of
syntax and semantics. The languages in XTREME are selected to maximize language diversity, coverage in existing tasks,
and availability of training data. Among these are many under-studied languages, such as the Dravidian languages Tamil
(spoken in southern India, Sri Lanka, and Singapore), Telugu and Malayalam (spoken mainly in southern India), and the
Niger-Congo languages Swahili and Yoruba, spoken in Africa.

• Лицензия : Нет известной лицензии.
• Версия : 1.0.0
• Расколы :

• Функции :

{
    "source_sentence": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "target_sentence": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "source_lang": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "target_lang": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

tatoeba.ell

Используйте следующую команду, чтобы загрузить этот набор данных в TFDS:

ds = tfds.load('huggingface:xtreme/tatoeba.ell')

• Описание :

his data is extracted from the Tatoeba corpus, dated Saturday 2018/11/17.
For each languages, we have selected 1000 English sentences and their translations, if available. Please check
this paper for a description of the languages, their families and scripts as well as baseline results.
Please note that the English sentences are not identical for all language pairs. This means that the results are
not directly comparable across languages. In particular, the sentences tend to have less variety for several
low-resource languages, e.g. "Tom needed water", "Tom needs water", "Tom is getting water", ...

The Cross-lingual TRansfer Evaluation of Multilingual Encoders (XTREME) benchmark is a benchmark for the evaluation of
the cross-lingual generalization ability of pre-trained multilingual models. It covers 40 typologically diverse languages
(spanning 12 language families) and includes nine tasks that collectively require reasoning about different levels of
syntax and semantics. The languages in XTREME are selected to maximize language diversity, coverage in existing tasks,
and availability of training data. Among these are many under-studied languages, such as the Dravidian languages Tamil
(spoken in southern India, Sri Lanka, and Singapore), Telugu and Malayalam (spoken mainly in southern India), and the
Niger-Congo languages Swahili and Yoruba, spoken in Africa.

• Лицензия : Нет известной лицензии.
• Версия : 1.0.0
• Расколы :

• Функции :

{
    "source_sentence": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "target_sentence": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "source_lang": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "target_lang": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

tatoeba.est

Используйте следующую команду, чтобы загрузить этот набор данных в TFDS:

ds = tfds.load('huggingface:xtreme/tatoeba.est')

• Описание :

his data is extracted from the Tatoeba corpus, dated Saturday 2018/11/17.
For each languages, we have selected 1000 English sentences and their translations, if available. Please check
this paper for a description of the languages, their families and scripts as well as baseline results.
Please note that the English sentences are not identical for all language pairs. This means that the results are
not directly comparable across languages. In particular, the sentences tend to have less variety for several
low-resource languages, e.g. "Tom needed water", "Tom needs water", "Tom is getting water", ...

The Cross-lingual TRansfer Evaluation of Multilingual Encoders (XTREME) benchmark is a benchmark for the evaluation of
the cross-lingual generalization ability of pre-trained multilingual models. It covers 40 typologically diverse languages
(spanning 12 language families) and includes nine tasks that collectively require reasoning about different levels of
syntax and semantics. The languages in XTREME are selected to maximize language diversity, coverage in existing tasks,
and availability of training data. Among these are many under-studied languages, such as the Dravidian languages Tamil
(spoken in southern India, Sri Lanka, and Singapore), Telugu and Malayalam (spoken mainly in southern India), and the
Niger-Congo languages Swahili and Yoruba, spoken in Africa.

• Лицензия : Нет известной лицензии.
• Версия : 1.0.0
• Расколы :

• Функции :

{
    "source_sentence": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "target_sentence": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "source_lang": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "target_lang": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

tatoeba.eus

Используйте следующую команду, чтобы загрузить этот набор данных в TFDS:

ds = tfds.load('huggingface:xtreme/tatoeba.eus')

• Описание :

his data is extracted from the Tatoeba corpus, dated Saturday 2018/11/17.
For each languages, we have selected 1000 English sentences and their translations, if available. Please check
this paper for a description of the languages, their families and scripts as well as baseline results.
Please note that the English sentences are not identical for all language pairs. This means that the results are
not directly comparable across languages. In particular, the sentences tend to have less variety for several
low-resource languages, e.g. "Tom needed water", "Tom needs water", "Tom is getting water", ...

The Cross-lingual TRansfer Evaluation of Multilingual Encoders (XTREME) benchmark is a benchmark for the evaluation of
the cross-lingual generalization ability of pre-trained multilingual models. It covers 40 typologically diverse languages
(spanning 12 language families) and includes nine tasks that collectively require reasoning about different levels of
syntax and semantics. The languages in XTREME are selected to maximize language diversity, coverage in existing tasks,
and availability of training data. Among these are many under-studied languages, such as the Dravidian languages Tamil
(spoken in southern India, Sri Lanka, and Singapore), Telugu and Malayalam (spoken mainly in southern India), and the
Niger-Congo languages Swahili and Yoruba, spoken in Africa.

• Лицензия : Нет известной лицензии.
• Версия : 1.0.0
• Расколы :

• Функции :

{
    "source_sentence": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "target_sentence": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "source_lang": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "target_lang": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

tatoeba.fin

Используйте следующую команду, чтобы загрузить этот набор данных в TFDS:

ds = tfds.load('huggingface:xtreme/tatoeba.fin')

• Описание :

his data is extracted from the Tatoeba corpus, dated Saturday 2018/11/17.
For each languages, we have selected 1000 English sentences and their translations, if available. Please check
this paper for a description of the languages, their families and scripts as well as baseline results.
Please note that the English sentences are not identical for all language pairs. This means that the results are
not directly comparable across languages. In particular, the sentences tend to have less variety for several
low-resource languages, e.g. "Tom needed water", "Tom needs water", "Tom is getting water", ...

The Cross-lingual TRansfer Evaluation of Multilingual Encoders (XTREME) benchmark is a benchmark for the evaluation of
the cross-lingual generalization ability of pre-trained multilingual models. It covers 40 typologically diverse languages
(spanning 12 language families) and includes nine tasks that collectively require reasoning about different levels of
syntax and semantics. The languages in XTREME are selected to maximize language diversity, coverage in existing tasks,
and availability of training data. Among these are many under-studied languages, such as the Dravidian languages Tamil
(spoken in southern India, Sri Lanka, and Singapore), Telugu and Malayalam (spoken mainly in southern India), and the
Niger-Congo languages Swahili and Yoruba, spoken in Africa.

• Лицензия : Нет известной лицензии.
• Версия : 1.0.0
• Расколы :

• Функции :

{
    "source_sentence": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "target_sentence": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "source_lang": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "target_lang": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

tatoeba.fra

Используйте следующую команду, чтобы загрузить этот набор данных в TFDS:

ds = tfds.load('huggingface:xtreme/tatoeba.fra')

• Описание :

his data is extracted from the Tatoeba corpus, dated Saturday 2018/11/17.
For each languages, we have selected 1000 English sentences and their translations, if available. Please check
this paper for a description of the languages, their families and scripts as well as baseline results.
Please note that the English sentences are not identical for all language pairs. This means that the results are
not directly comparable across languages. In particular, the sentences tend to have less variety for several
low-resource languages, e.g. "Tom needed water", "Tom needs water", "Tom is getting water", ...

The Cross-lingual TRansfer Evaluation of Multilingual Encoders (XTREME) benchmark is a benchmark for the evaluation of
the cross-lingual generalization ability of pre-trained multilingual models. It covers 40 typologically diverse languages
(spanning 12 language families) and includes nine tasks that collectively require reasoning about different levels of
syntax and semantics. The languages in XTREME are selected to maximize language diversity, coverage in existing tasks,
and availability of training data. Among these are many under-studied languages, such as the Dravidian languages Tamil
(spoken in southern India, Sri Lanka, and Singapore), Telugu and Malayalam (spoken mainly in southern India), and the
Niger-Congo languages Swahili and Yoruba, spoken in Africa.

• Лицензия : Нет известной лицензии.
• Версия : 1.0.0
• Расколы :

• Функции :

{
    "source_sentence": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "target_sentence": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "source_lang": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "target_lang": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

tatoeba.heb

Используйте следующую команду, чтобы загрузить этот набор данных в TFDS:

ds = tfds.load('huggingface:xtreme/tatoeba.heb')

• Описание :

his data is extracted from the Tatoeba corpus, dated Saturday 2018/11/17.
For each languages, we have selected 1000 English sentences and their translations, if available. Please check
this paper for a description of the languages, their families and scripts as well as baseline results.
Please note that the English sentences are not identical for all language pairs. This means that the results are
not directly comparable across languages. In particular, the sentences tend to have less variety for several
low-resource languages, e.g. "Tom needed water", "Tom needs water", "Tom is getting water", ...

The Cross-lingual TRansfer Evaluation of Multilingual Encoders (XTREME) benchmark is a benchmark for the evaluation of
the cross-lingual generalization ability of pre-trained multilingual models. It covers 40 typologically diverse languages
(spanning 12 language families) and includes nine tasks that collectively require reasoning about different levels of
syntax and semantics. The languages in XTREME are selected to maximize language diversity, coverage in existing tasks,
and availability of training data. Among these are many under-studied languages, such as the Dravidian languages Tamil
(spoken in southern India, Sri Lanka, and Singapore), Telugu and Malayalam (spoken mainly in southern India), and the
Niger-Congo languages Swahili and Yoruba, spoken in Africa.

• Лицензия : Нет известной лицензии.
• Версия : 1.0.0
• Расколы :

• Функции :

{
    "source_sentence": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "target_sentence": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "source_lang": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "target_lang": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

tatoeba.hin

Используйте следующую команду, чтобы загрузить этот набор данных в TFDS:

ds = tfds.load('huggingface:xtreme/tatoeba.hin')

• Описание :

his data is extracted from the Tatoeba corpus, dated Saturday 2018/11/17.
For each languages, we have selected 1000 English sentences and their translations, if available. Please check
this paper for a description of the languages, their families and scripts as well as baseline results.
Please note that the English sentences are not identical for all language pairs. This means that the results are
not directly comparable across languages. In particular, the sentences tend to have less variety for several
low-resource languages, e.g. "Tom needed water", "Tom needs water", "Tom is getting water", ...

The Cross-lingual TRansfer Evaluation of Multilingual Encoders (XTREME) benchmark is a benchmark for the evaluation of
the cross-lingual generalization ability of pre-trained multilingual models. It covers 40 typologically diverse languages
(spanning 12 language families) and includes nine tasks that collectively require reasoning about different levels of
syntax and semantics. The languages in XTREME are selected to maximize language diversity, coverage in existing tasks,
and availability of training data. Among these are many under-studied languages, such as the Dravidian languages Tamil
(spoken in southern India, Sri Lanka, and Singapore), Telugu and Malayalam (spoken mainly in southern India), and the
Niger-Congo languages Swahili and Yoruba, spoken in Africa.

• Лицензия : Нет известной лицензии.
• Версия : 1.0.0
• Расколы :

• Функции :

{
    "source_sentence": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "target_sentence": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "source_lang": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "target_lang": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

tatoeba.hun

Используйте следующую команду, чтобы загрузить этот набор данных в TFDS:

ds = tfds.load('huggingface:xtreme/tatoeba.hun')

• Описание :

his data is extracted from the Tatoeba corpus, dated Saturday 2018/11/17.
For each languages, we have selected 1000 English sentences and their translations, if available. Please check
this paper for a description of the languages, their families and scripts as well as baseline results.
Please note that the English sentences are not identical for all language pairs. This means that the results are
not directly comparable across languages. In particular, the sentences tend to have less variety for several
low-resource languages, e.g. "Tom needed water", "Tom needs water", "Tom is getting water", ...

The Cross-lingual TRansfer Evaluation of Multilingual Encoders (XTREME) benchmark is a benchmark for the evaluation of
the cross-lingual generalization ability of pre-trained multilingual models. It covers 40 typologically diverse languages
(spanning 12 language families) and includes nine tasks that collectively require reasoning about different levels of
syntax and semantics. The languages in XTREME are selected to maximize language diversity, coverage in existing tasks,
and availability of training data. Among these are many under-studied languages, such as the Dravidian languages Tamil
(spoken in southern India, Sri Lanka, and Singapore), Telugu and Malayalam (spoken mainly in southern India), and the
Niger-Congo languages Swahili and Yoruba, spoken in Africa.

• Лицензия : Нет известной лицензии.
• Версия : 1.0.0
• Расколы :

• Функции :

{
    "source_sentence": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "target_sentence": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "source_lang": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "target_lang": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

tatoeba.ind

Используйте следующую команду, чтобы загрузить этот набор данных в TFDS:

ds = tfds.load('huggingface:xtreme/tatoeba.ind')

• Описание :

his data is extracted from the Tatoeba corpus, dated Saturday 2018/11/17.
For each languages, we have selected 1000 English sentences and their translations, if available. Please check
this paper for a description of the languages, their families and scripts as well as baseline results.
Please note that the English sentences are not identical for all language pairs. This means that the results are
not directly comparable across languages. In particular, the sentences tend to have less variety for several
low-resource languages, e.g. "Tom needed water", "Tom needs water", "Tom is getting water", ...

The Cross-lingual TRansfer Evaluation of Multilingual Encoders (XTREME) benchmark is a benchmark for the evaluation of
the cross-lingual generalization ability of pre-trained multilingual models. It covers 40 typologically diverse languages
(spanning 12 language families) and includes nine tasks that collectively require reasoning about different levels of
syntax and semantics. The languages in XTREME are selected to maximize language diversity, coverage in existing tasks,
and availability of training data. Among these are many under-studied languages, such as the Dravidian languages Tamil
(spoken in southern India, Sri Lanka, and Singapore), Telugu and Malayalam (spoken mainly in southern India), and the
Niger-Congo languages Swahili and Yoruba, spoken in Africa.

• Лицензия : Нет известной лицензии.
• Версия : 1.0.0
• Расколы :

• Функции :

{
    "source_sentence": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "target_sentence": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "source_lang": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "target_lang": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

tatoeba.ita

Используйте следующую команду, чтобы загрузить этот набор данных в TFDS:

ds = tfds.load('huggingface:xtreme/tatoeba.ita')

• Описание :

his data is extracted from the Tatoeba corpus, dated Saturday 2018/11/17.
For each languages, we have selected 1000 English sentences and their translations, if available. Please check
this paper for a description of the languages, their families and scripts as well as baseline results.
Please note that the English sentences are not identical for all language pairs. This means that the results are
not directly comparable across languages. In particular, the sentences tend to have less variety for several
low-resource languages, e.g. "Tom needed water", "Tom needs water", "Tom is getting water", ...

The Cross-lingual TRansfer Evaluation of Multilingual Encoders (XTREME) benchmark is a benchmark for the evaluation of
the cross-lingual generalization ability of pre-trained multilingual models. It covers 40 typologically diverse languages
(spanning 12 language families) and includes nine tasks that collectively require reasoning about different levels of
syntax and semantics. The languages in XTREME are selected to maximize language diversity, coverage in existing tasks,
and availability of training data. Among these are many under-studied languages, such as the Dravidian languages Tamil
(spoken in southern India, Sri Lanka, and Singapore), Telugu and Malayalam (spoken mainly in southern India), and the
Niger-Congo languages Swahili and Yoruba, spoken in Africa.

• Лицензия : Нет известной лицензии.
• Версия : 1.0.0
• Расколы :

• Функции :

{
    "source_sentence": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "target_sentence": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "source_lang": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "target_lang": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

tatoeba.jav

Используйте следующую команду, чтобы загрузить этот набор данных в TFDS:

ds = tfds.load('huggingface:xtreme/tatoeba.jav')

• Описание :

his data is extracted from the Tatoeba corpus, dated Saturday 2018/11/17.
For each languages, we have selected 1000 English sentences and their translations, if available. Please check
this paper for a description of the languages, their families and scripts as well as baseline results.
Please note that the English sentences are not identical for all language pairs. This means that the results are
not directly comparable across languages. In particular, the sentences tend to have less variety for several
low-resource languages, e.g. "Tom needed water", "Tom needs water", "Tom is getting water", ...

The Cross-lingual TRansfer Evaluation of Multilingual Encoders (XTREME) benchmark is a benchmark for the evaluation of
the cross-lingual generalization ability of pre-trained multilingual models. It covers 40 typologically diverse languages
(spanning 12 language families) and includes nine tasks that collectively require reasoning about different levels of
syntax and semantics. The languages in XTREME are selected to maximize language diversity, coverage in existing tasks,
and availability of training data. Among these are many under-studied languages, such as the Dravidian languages Tamil
(spoken in southern India, Sri Lanka, and Singapore), Telugu and Malayalam (spoken mainly in southern India), and the
Niger-Congo languages Swahili and Yoruba, spoken in Africa.

• Лицензия : Нет известной лицензии.
• Версия : 1.0.0
• Расколы :

• Функции :

{
    "source_sentence": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "target_sentence": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "source_lang": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "target_lang": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

tatoeba.jpn

Используйте следующую команду, чтобы загрузить этот набор данных в TFDS:

ds = tfds.load('huggingface:xtreme/tatoeba.jpn')

• Описание :

his data is extracted from the Tatoeba corpus, dated Saturday 2018/11/17.
For each languages, we have selected 1000 English sentences and their translations, if available. Please check
this paper for a description of the languages, their families and scripts as well as baseline results.
Please note that the English sentences are not identical for all language pairs. This means that the results are
not directly comparable across languages. In particular, the sentences tend to have less variety for several
low-resource languages, e.g. "Tom needed water", "Tom needs water", "Tom is getting water", ...

The Cross-lingual TRansfer Evaluation of Multilingual Encoders (XTREME) benchmark is a benchmark for the evaluation of
the cross-lingual generalization ability of pre-trained multilingual models. It covers 40 typologically diverse languages
(spanning 12 language families) and includes nine tasks that collectively require reasoning about different levels of
syntax and semantics. The languages in XTREME are selected to maximize language diversity, coverage in existing tasks,
and availability of training data. Among these are many under-studied languages, such as the Dravidian languages Tamil
(spoken in southern India, Sri Lanka, and Singapore), Telugu and Malayalam (spoken mainly in southern India), and the
Niger-Congo languages Swahili and Yoruba, spoken in Africa.

• Лицензия : Нет известной лицензии.
• Версия : 1.0.0
• Расколы :

• Функции :

{
    "source_sentence": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "target_sentence": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "source_lang": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "target_lang": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

tatoeba.kat

Используйте следующую команду, чтобы загрузить этот набор данных в TFDS:

ds = tfds.load('huggingface:xtreme/tatoeba.kat')

• Описание :

his data is extracted from the Tatoeba corpus, dated Saturday 2018/11/17.
For each languages, we have selected 1000 English sentences and their translations, if available. Please check
this paper for a description of the languages, their families and scripts as well as baseline results.
Please note that the English sentences are not identical for all language pairs. This means that the results are
not directly comparable across languages. In particular, the sentences tend to have less variety for several
low-resource languages, e.g. "Tom needed water", "Tom needs water", "Tom is getting water", ...

The Cross-lingual TRansfer Evaluation of Multilingual Encoders (XTREME) benchmark is a benchmark for the evaluation of
the cross-lingual generalization ability of pre-trained multilingual models. It covers 40 typologically diverse languages
(spanning 12 language families) and includes nine tasks that collectively require reasoning about different levels of
syntax and semantics. The languages in XTREME are selected to maximize language diversity, coverage in existing tasks,
and availability of training data. Among these are many under-studied languages, such as the Dravidian languages Tamil
(spoken in southern India, Sri Lanka, and Singapore), Telugu and Malayalam (spoken mainly in southern India), and the
Niger-Congo languages Swahili and Yoruba, spoken in Africa.

• Лицензия : Нет известной лицензии.
• Версия : 1.0.0
• Расколы :

• Функции :

{
    "source_sentence": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "target_sentence": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "source_lang": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "target_lang": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

tatoeba.kaz

Используйте следующую команду, чтобы загрузить этот набор данных в TFDS:

ds = tfds.load('huggingface:xtreme/tatoeba.kaz')

• Описание :

his data is extracted from the Tatoeba corpus, dated Saturday 2018/11/17.
For each languages, we have selected 1000 English sentences and their translations, if available. Please check
this paper for a description of the languages, their families and scripts as well as baseline results.
Please note that the English sentences are not identical for all language pairs. This means that the results are
not directly comparable across languages. In particular, the sentences tend to have less variety for several
low-resource languages, e.g. "Tom needed water", "Tom needs water", "Tom is getting water", ...

The Cross-lingual TRansfer Evaluation of Multilingual Encoders (XTREME) benchmark is a benchmark for the evaluation of
the cross-lingual generalization ability of pre-trained multilingual models. It covers 40 typologically diverse languages
(spanning 12 language families) and includes nine tasks that collectively require reasoning about different levels of
syntax and semantics. The languages in XTREME are selected to maximize language diversity, coverage in existing tasks,
and availability of training data. Among these are many under-studied languages, such as the Dravidian languages Tamil
(spoken in southern India, Sri Lanka, and Singapore), Telugu and Malayalam (spoken mainly in southern India), and the
Niger-Congo languages Swahili and Yoruba, spoken in Africa.

• Лицензия : Нет известной лицензии.
• Версия : 1.0.0
• Расколы :

• Функции :

{
    "source_sentence": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "target_sentence": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "source_lang": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "target_lang": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

tatoeba.kor

Используйте следующую команду, чтобы загрузить этот набор данных в TFDS:

ds = tfds.load('huggingface:xtreme/tatoeba.kor')

• Описание :

his data is extracted from the Tatoeba corpus, dated Saturday 2018/11/17.
For each languages, we have selected 1000 English sentences and their translations, if available. Please check
this paper for a description of the languages, their families and scripts as well as baseline results.
Please note that the English sentences are not identical for all language pairs. This means that the results are
not directly comparable across languages. In particular, the sentences tend to have less variety for several
low-resource languages, e.g. "Tom needed water", "Tom needs water", "Tom is getting water", ...

The Cross-lingual TRansfer Evaluation of Multilingual Encoders (XTREME) benchmark is a benchmark for the evaluation of
the cross-lingual generalization ability of pre-trained multilingual models. It covers 40 typologically diverse languages
(spanning 12 language families) and includes nine tasks that collectively require reasoning about different levels of
syntax and semantics. The languages in XTREME are selected to maximize language diversity, coverage in existing tasks,
and availability of training data. Among these are many under-studied languages, such as the Dravidian languages Tamil
(spoken in southern India, Sri Lanka, and Singapore), Telugu and Malayalam (spoken mainly in southern India), and the
Niger-Congo languages Swahili and Yoruba, spoken in Africa.

• Лицензия : Нет известной лицензии.
• Версия : 1.0.0
• Расколы :

• Функции :

{
    "source_sentence": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "target_sentence": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "source_lang": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "target_lang": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

tatoeba.mal

Используйте следующую команду, чтобы загрузить этот набор данных в TFDS:

ds = tfds.load('huggingface:xtreme/tatoeba.mal')

• Описание :

his data is extracted from the Tatoeba corpus, dated Saturday 2018/11/17.
For each languages, we have selected 1000 English sentences and their translations, if available. Please check
this paper for a description of the languages, their families and scripts as well as baseline results.
Please note that the English sentences are not identical for all language pairs. This means that the results are
not directly comparable across languages. In particular, the sentences tend to have less variety for several
low-resource languages, e.g. "Tom needed water", "Tom needs water", "Tom is getting water", ...

The Cross-lingual TRansfer Evaluation of Multilingual Encoders (XTREME) benchmark is a benchmark for the evaluation of
the cross-lingual generalization ability of pre-trained multilingual models. It covers 40 typologically diverse languages
(spanning 12 language families) and includes nine tasks that collectively require reasoning about different levels of
syntax and semantics. The languages in XTREME are selected to maximize language diversity, coverage in existing tasks,
and availability of training data. Among these are many under-studied languages, such as the Dravidian languages Tamil
(spoken in southern India, Sri Lanka, and Singapore), Telugu and Malayalam (spoken mainly in southern India), and the
Niger-Congo languages Swahili and Yoruba, spoken in Africa.

• Лицензия : Нет известной лицензии.
• Версия : 1.0.0
• Расколы :

• Функции :

{
    "source_sentence": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "target_sentence": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "source_lang": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "target_lang": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

tatoeba.mar

Используйте следующую команду, чтобы загрузить этот набор данных в TFDS:

ds = tfds.load('huggingface:xtreme/tatoeba.mar')

• Описание :

his data is extracted from the Tatoeba corpus, dated Saturday 2018/11/17.
For each languages, we have selected 1000 English sentences and their translations, if available. Please check
this paper for a description of the languages, their families and scripts as well as baseline results.
Please note that the English sentences are not identical for all language pairs. This means that the results are
not directly comparable across languages. In particular, the sentences tend to have less variety for several
low-resource languages, e.g. "Tom needed water", "Tom needs water", "Tom is getting water", ...

The Cross-lingual TRansfer Evaluation of Multilingual Encoders (XTREME) benchmark is a benchmark for the evaluation of
the cross-lingual generalization ability of pre-trained multilingual models. It covers 40 typologically diverse languages
(spanning 12 language families) and includes nine tasks that collectively require reasoning about different levels of
syntax and semantics. The languages in XTREME are selected to maximize language diversity, coverage in existing tasks,
and availability of training data. Among these are many under-studied languages, such as the Dravidian languages Tamil
(spoken in southern India, Sri Lanka, and Singapore), Telugu and Malayalam (spoken mainly in southern India), and the
Niger-Congo languages Swahili and Yoruba, spoken in Africa.

• Лицензия : Нет известной лицензии.
• Версия : 1.0.0
• Расколы :

• Функции :

{
    "source_sentence": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "target_sentence": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "source_lang": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "target_lang": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

tatoeba.nld

Используйте следующую команду, чтобы загрузить этот набор данных в TFDS:

ds = tfds.load('huggingface:xtreme/tatoeba.nld')

• Описание :

his data is extracted from the Tatoeba corpus, dated Saturday 2018/11/17.
For each languages, we have selected 1000 English sentences and their translations, if available. Please check
this paper for a description of the languages, their families and scripts as well as baseline results.
Please note that the English sentences are not identical for all language pairs. This means that the results are
not directly comparable across languages. In particular, the sentences tend to have less variety for several
low-resource languages, e.g. "Tom needed water", "Tom needs water", "Tom is getting water", ...

The Cross-lingual TRansfer Evaluation of Multilingual Encoders (XTREME) benchmark is a benchmark for the evaluation of
the cross-lingual generalization ability of pre-trained multilingual models. It covers 40 typologically diverse languages
(spanning 12 language families) and includes nine tasks that collectively require reasoning about different levels of
syntax and semantics. The languages in XTREME are selected to maximize language diversity, coverage in existing tasks,
and availability of training data. Among these are many under-studied languages, such as the Dravidian languages Tamil
(spoken in southern India, Sri Lanka, and Singapore), Telugu and Malayalam (spoken mainly in southern India), and the
Niger-Congo languages Swahili and Yoruba, spoken in Africa.

• Лицензия : Нет известной лицензии.
• Версия : 1.0.0
• Расколы :

• Функции :

{
    "source_sentence": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "target_sentence": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "source_lang": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "target_lang": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

tatoeba.pes

Используйте следующую команду, чтобы загрузить этот набор данных в TFDS:

ds = tfds.load('huggingface:xtreme/tatoeba.pes')

• Описание :

his data is extracted from the Tatoeba corpus, dated Saturday 2018/11/17.
For each languages, we have selected 1000 English sentences and their translations, if available. Please check
this paper for a description of the languages, their families and scripts as well as baseline results.
Please note that the English sentences are not identical for all language pairs. This means that the results are
not directly comparable across languages. In particular, the sentences tend to have less variety for several
low-resource languages, e.g. "Tom needed water", "Tom needs water", "Tom is getting water", ...

The Cross-lingual TRansfer Evaluation of Multilingual Encoders (XTREME) benchmark is a benchmark for the evaluation of
the cross-lingual generalization ability of pre-trained multilingual models. It covers 40 typologically diverse languages
(spanning 12 language families) and includes nine tasks that collectively require reasoning about different levels of
syntax and semantics. The languages in XTREME are selected to maximize language diversity, coverage in existing tasks,
and availability of training data. Among these are many under-studied languages, such as the Dravidian languages Tamil
(spoken in southern India, Sri Lanka, and Singapore), Telugu and Malayalam (spoken mainly in southern India), and the
Niger-Congo languages Swahili and Yoruba, spoken in Africa.

• Лицензия : Нет известной лицензии.
• Версия : 1.0.0
• Расколы :

• Функции :

{
    "source_sentence": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "target_sentence": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "source_lang": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "target_lang": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

tatoeba.por

Используйте следующую команду, чтобы загрузить этот набор данных в TFDS:

ds = tfds.load('huggingface:xtreme/tatoeba.por')

• Описание :

his data is extracted from the Tatoeba corpus, dated Saturday 2018/11/17.
For each languages, we have selected 1000 English sentences and their translations, if available. Please check
this paper for a description of the languages, their families and scripts as well as baseline results.
Please note that the English sentences are not identical for all language pairs. This means that the results are
not directly comparable across languages. In particular, the sentences tend to have less variety for several
low-resource languages, e.g. "Tom needed water", "Tom needs water", "Tom is getting water", ...

The Cross-lingual TRansfer Evaluation of Multilingual Encoders (XTREME) benchmark is a benchmark for the evaluation of
the cross-lingual generalization ability of pre-trained multilingual models. It covers 40 typologically diverse languages
(spanning 12 language families) and includes nine tasks that collectively require reasoning about different levels of
syntax and semantics. The languages in XTREME are selected to maximize language diversity, coverage in existing tasks,
and availability of training data. Among these are many under-studied languages, such as the Dravidian languages Tamil
(spoken in southern India, Sri Lanka, and Singapore), Telugu and Malayalam (spoken mainly in southern India), and the
Niger-Congo languages Swahili and Yoruba, spoken in Africa.

• Лицензия : Нет известной лицензии.
• Версия : 1.0.0
• Расколы :

• Функции :

{
    "source_sentence": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "target_sentence": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "source_lang": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "target_lang": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

tatoeba.rus

Используйте следующую команду, чтобы загрузить этот набор данных в TFDS:

ds = tfds.load('huggingface:xtreme/tatoeba.rus')

• Описание :

his data is extracted from the Tatoeba corpus, dated Saturday 2018/11/17.
For each languages, we have selected 1000 English sentences and their translations, if available. Please check
this paper for a description of the languages, their families and scripts as well as baseline results.
Please note that the English sentences are not identical for all language pairs. This means that the results are
not directly comparable across languages. In particular, the sentences tend to have less variety for several
low-resource languages, e.g. "Tom needed water", "Tom needs water", "Tom is getting water", ...

The Cross-lingual TRansfer Evaluation of Multilingual Encoders (XTREME) benchmark is a benchmark for the evaluation of
the cross-lingual generalization ability of pre-trained multilingual models. It covers 40 typologically diverse languages
(spanning 12 language families) and includes nine tasks that collectively require reasoning about different levels of
syntax and semantics. The languages in XTREME are selected to maximize language diversity, coverage in existing tasks,
and availability of training data. Among these are many under-studied languages, such as the Dravidian languages Tamil
(spoken in southern India, Sri Lanka, and Singapore), Telugu and Malayalam (spoken mainly in southern India), and the
Niger-Congo languages Swahili and Yoruba, spoken in Africa.

• Лицензия : Нет известной лицензии.
• Версия : 1.0.0
• Расколы :

• Функции :

{
    "source_sentence": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "target_sentence": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "source_lang": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "target_lang": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

tatoeba.spa

Используйте следующую команду, чтобы загрузить этот набор данных в TFDS:

ds = tfds.load('huggingface:xtreme/tatoeba.spa')

• Описание :

his data is extracted from the Tatoeba corpus, dated Saturday 2018/11/17.
For each languages, we have selected 1000 English sentences and their translations, if available. Please check
this paper for a description of the languages, their families and scripts as well as baseline results.
Please note that the English sentences are not identical for all language pairs. This means that the results are
not directly comparable across languages. In particular, the sentences tend to have less variety for several
low-resource languages, e.g. "Tom needed water", "Tom needs water", "Tom is getting water", ...

The Cross-lingual TRansfer Evaluation of Multilingual Encoders (XTREME) benchmark is a benchmark for the evaluation of
the cross-lingual generalization ability of pre-trained multilingual models. It covers 40 typologically diverse languages
(spanning 12 language families) and includes nine tasks that collectively require reasoning about different levels of
syntax and semantics. The languages in XTREME are selected to maximize language diversity, coverage in existing tasks,
and availability of training data. Among these are many under-studied languages, such as the Dravidian languages Tamil
(spoken in southern India, Sri Lanka, and Singapore), Telugu and Malayalam (spoken mainly in southern India), and the
Niger-Congo languages Swahili and Yoruba, spoken in Africa.

• Лицензия : Нет известной лицензии.
• Версия : 1.0.0
• Расколы :

• Функции :

{
    "source_sentence": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "target_sentence": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "source_lang": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "target_lang": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

tatoeba.swh

Используйте следующую команду, чтобы загрузить этот набор данных в TFDS:

ds = tfds.load('huggingface:xtreme/tatoeba.swh')

• Описание :

his data is extracted from the Tatoeba corpus, dated Saturday 2018/11/17.
For each languages, we have selected 1000 English sentences and their translations, if available. Please check
this paper for a description of the languages, their families and scripts as well as baseline results.
Please note that the English sentences are not identical for all language pairs. This means that the results are
not directly comparable across languages. In particular, the sentences tend to have less variety for several
low-resource languages, e.g. "Tom needed water", "Tom needs water", "Tom is getting water", ...

The Cross-lingual TRansfer Evaluation of Multilingual Encoders (XTREME) benchmark is a benchmark for the evaluation of
the cross-lingual generalization ability of pre-trained multilingual models. It covers 40 typologically diverse languages
(spanning 12 language families) and includes nine tasks that collectively require reasoning about different levels of
syntax and semantics. The languages in XTREME are selected to maximize language diversity, coverage in existing tasks,
and availability of training data. Among these are many under-studied languages, such as the Dravidian languages Tamil
(spoken in southern India, Sri Lanka, and Singapore), Telugu and Malayalam (spoken mainly in southern India), and the
Niger-Congo languages Swahili and Yoruba, spoken in Africa.

• Лицензия : Нет известной лицензии.
• Версия : 1.0.0
• Расколы :

• Функции :

{
    "source_sentence": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "target_sentence": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "source_lang": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "target_lang": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

tatoeba.tam

Используйте следующую команду, чтобы загрузить этот набор данных в TFDS:

ds = tfds.load('huggingface:xtreme/tatoeba.tam')

• Описание :

his data is extracted from the Tatoeba corpus, dated Saturday 2018/11/17.
For each languages, we have selected 1000 English sentences and their translations, if available. Please check
this paper for a description of the languages, their families and scripts as well as baseline results.
Please note that the English sentences are not identical for all language pairs. This means that the results are
not directly comparable across languages. In particular, the sentences tend to have less variety for several
low-resource languages, e.g. "Tom needed water", "Tom needs water", "Tom is getting water", ...

The Cross-lingual TRansfer Evaluation of Multilingual Encoders (XTREME) benchmark is a benchmark for the evaluation of
the cross-lingual generalization ability of pre-trained multilingual models. It covers 40 typologically diverse languages
(spanning 12 language families) and includes nine tasks that collectively require reasoning about different levels of
syntax and semantics. The languages in XTREME are selected to maximize language diversity, coverage in existing tasks,
and availability of training data. Among these are many under-studied languages, such as the Dravidian languages Tamil
(spoken in southern India, Sri Lanka, and Singapore), Telugu and Malayalam (spoken mainly in southern India), and the
Niger-Congo languages Swahili and Yoruba, spoken in Africa.

• Лицензия : Нет известной лицензии.
• Версия : 1.0.0
• Расколы :

• Функции :

{
    "source_sentence": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "target_sentence": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "source_lang": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "target_lang": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

tatoeba.tel

Используйте следующую команду, чтобы загрузить этот набор данных в TFDS:

ds = tfds.load('huggingface:xtreme/tatoeba.tel')

• Описание :

his data is extracted from the Tatoeba corpus, dated Saturday 2018/11/17.
For each languages, we have selected 1000 English sentences and their translations, if available. Please check
this paper for a description of the languages, their families and scripts as well as baseline results.
Please note that the English sentences are not identical for all language pairs. This means that the results are
not directly comparable across languages. In particular, the sentences tend to have less variety for several
low-resource languages, e.g. "Tom needed water", "Tom needs water", "Tom is getting water", ...

The Cross-lingual TRansfer Evaluation of Multilingual Encoders (XTREME) benchmark is a benchmark for the evaluation of
the cross-lingual generalization ability of pre-trained multilingual models. It covers 40 typologically diverse languages
(spanning 12 language families) and includes nine tasks that collectively require reasoning about different levels of
syntax and semantics. The languages in XTREME are selected to maximize language diversity, coverage in existing tasks,
and availability of training data. Among these are many under-studied languages, such as the Dravidian languages Tamil
(spoken in southern India, Sri Lanka, and Singapore), Telugu and Malayalam (spoken mainly in southern India), and the
Niger-Congo languages Swahili and Yoruba, spoken in Africa.

• Лицензия : Нет известной лицензии.
• Версия : 1.0.0
• Расколы :

• Функции :

{
    "source_sentence": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "target_sentence": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "source_lang": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "target_lang": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

tatoeba.tgl

Используйте следующую команду, чтобы загрузить этот набор данных в TFDS:

ds = tfds.load('huggingface:xtreme/tatoeba.tgl')

• Описание :

his data is extracted from the Tatoeba corpus, dated Saturday 2018/11/17.
For each languages, we have selected 1000 English sentences and their translations, if available. Please check
this paper for a description of the languages, their families and scripts as well as baseline results.
Please note that the English sentences are not identical for all language pairs. This means that the results are
not directly comparable across languages. In particular, the sentences tend to have less variety for several
low-resource languages, e.g. "Tom needed water", "Tom needs water", "Tom is getting water", ...

The Cross-lingual TRansfer Evaluation of Multilingual Encoders (XTREME) benchmark is a benchmark for the evaluation of
the cross-lingual generalization ability of pre-trained multilingual models. It covers 40 typologically diverse languages
(spanning 12 language families) and includes nine tasks that collectively require reasoning about different levels of
syntax and semantics. The languages in XTREME are selected to maximize language diversity, coverage in existing tasks,
and availability of training data. Among these are many under-studied languages, such as the Dravidian languages Tamil
(spoken in southern India, Sri Lanka, and Singapore), Telugu and Malayalam (spoken mainly in southern India), and the
Niger-Congo languages Swahili and Yoruba, spoken in Africa.

• Лицензия : Нет известной лицензии.
• Версия : 1.0.0
• Расколы :

• Функции :

{
    "source_sentence": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "target_sentence": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "source_lang": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "target_lang": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

tatoeba.tha

Используйте следующую команду, чтобы загрузить этот набор данных в TFDS:

ds = tfds.load('huggingface:xtreme/tatoeba.tha')

• Описание :

his data is extracted from the Tatoeba corpus, dated Saturday 2018/11/17.
For each languages, we have selected 1000 English sentences and their translations, if available. Please check
this paper for a description of the languages, their families and scripts as well as baseline results.
Please note that the English sentences are not identical for all language pairs. This means that the results are
not directly comparable across languages. In particular, the sentences tend to have less variety for several
low-resource languages, e.g. "Tom needed water", "Tom needs water", "Tom is getting water", ...

The Cross-lingual TRansfer Evaluation of Multilingual Encoders (XTREME) benchmark is a benchmark for the evaluation of
the cross-lingual generalization ability of pre-trained multilingual models. It covers 40 typologically diverse languages
(spanning 12 language families) and includes nine tasks that collectively require reasoning about different levels of
syntax and semantics. The languages in XTREME are selected to maximize language diversity, coverage in existing tasks,
and availability of training data. Among these are many under-studied languages, such as the Dravidian languages Tamil
(spoken in southern India, Sri Lanka, and Singapore), Telugu and Malayalam (spoken mainly in southern India), and the
Niger-Congo languages Swahili and Yoruba, spoken in Africa.

• Лицензия : Нет известной лицензии.
• Версия : 1.0.0
• Расколы :

• Функции :

{
    "source_sentence": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "target_sentence": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "source_lang": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "target_lang": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

tatoeba.tur

Используйте следующую команду, чтобы загрузить этот набор данных в TFDS:

ds = tfds.load('huggingface:xtreme/tatoeba.tur')

• Описание :

his data is extracted from the Tatoeba corpus, dated Saturday 2018/11/17.
For each languages, we have selected 1000 English sentences and their translations, if available. Please check
this paper for a description of the languages, their families and scripts as well as baseline results.
Please note that the English sentences are not identical for all language pairs. This means that the results are
not directly comparable across languages. In particular, the sentences tend to have less variety for several
low-resource languages, e.g. "Tom needed water", "Tom needs water", "Tom is getting water", ...

The Cross-lingual TRansfer Evaluation of Multilingual Encoders (XTREME) benchmark is a benchmark for the evaluation of
the cross-lingual generalization ability of pre-trained multilingual models. It covers 40 typologically diverse languages
(spanning 12 language families) and includes nine tasks that collectively require reasoning about different levels of
syntax and semantics. The languages in XTREME are selected to maximize language diversity, coverage in existing tasks,
and availability of training data. Among these are many under-studied languages, such as the Dravidian languages Tamil
(spoken in southern India, Sri Lanka, and Singapore), Telugu and Malayalam (spoken mainly in southern India), and the
Niger-Congo languages Swahili and Yoruba, spoken in Africa.

• Лицензия : Нет известной лицензии.
• Версия : 1.0.0
• Расколы :

• Функции :

{
    "source_sentence": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "target_sentence": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "source_lang": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "target_lang": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

tatoeba.urd

Используйте следующую команду, чтобы загрузить этот набор данных в TFDS:

ds = tfds.load('huggingface:xtreme/tatoeba.urd')

• Описание :

his data is extracted from the Tatoeba corpus, dated Saturday 2018/11/17.
For each languages, we have selected 1000 English sentences and their translations, if available. Please check
this paper for a description of the languages, their families and scripts as well as baseline results.
Please note that the English sentences are not identical for all language pairs. This means that the results are
not directly comparable across languages. In particular, the sentences tend to have less variety for several
low-resource languages, e.g. "Tom needed water", "Tom needs water", "Tom is getting water", ...

The Cross-lingual TRansfer Evaluation of Multilingual Encoders (XTREME) benchmark is a benchmark for the evaluation of
the cross-lingual generalization ability of pre-trained multilingual models. It covers 40 typologically diverse languages
(spanning 12 language families) and includes nine tasks that collectively require reasoning about different levels of
syntax and semantics. The languages in XTREME are selected to maximize language diversity, coverage in existing tasks,
and availability of training data. Among these are many under-studied languages, such as the Dravidian languages Tamil
(spoken in southern India, Sri Lanka, and Singapore), Telugu and Malayalam (spoken mainly in southern India), and the
Niger-Congo languages Swahili and Yoruba, spoken in Africa.

• Лицензия : Нет известной лицензии.
• Версия : 1.0.0
• Расколы :

• Функции :

{
    "source_sentence": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "target_sentence": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "source_lang": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "target_lang": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

tatoeba.vie

Используйте следующую команду, чтобы загрузить этот набор данных в TFDS:

ds = tfds.load('huggingface:xtreme/tatoeba.vie')

• Описание :

his data is extracted from the Tatoeba corpus, dated Saturday 2018/11/17.
For each languages, we have selected 1000 English sentences and their translations, if available. Please check
this paper for a description of the languages, their families and scripts as well as baseline results.
Please note that the English sentences are not identical for all language pairs. This means that the results are
not directly comparable across languages. In particular, the sentences tend to have less variety for several
low-resource languages, e.g. "Tom needed water", "Tom needs water", "Tom is getting water", ...

The Cross-lingual TRansfer Evaluation of Multilingual Encoders (XTREME) benchmark is a benchmark for the evaluation of
the cross-lingual generalization ability of pre-trained multilingual models. It covers 40 typologically diverse languages
(spanning 12 language families) and includes nine tasks that collectively require reasoning about different levels of
syntax and semantics. The languages in XTREME are selected to maximize language diversity, coverage in existing tasks,
and availability of training data. Among these are many under-studied languages, such as the Dravidian languages Tamil
(spoken in southern India, Sri Lanka, and Singapore), Telugu and Malayalam (spoken mainly in southern India), and the
Niger-Congo languages Swahili and Yoruba, spoken in Africa.

• Лицензия : Нет известной лицензии.
• Версия : 1.0.0
• Расколы :

• Функции :

{
    "source_sentence": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "target_sentence": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "source_lang": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "target_lang": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

udpos.Африкаанс

Используйте следующую команду, чтобы загрузить этот набор данных в TFDS:

ds = tfds.load('huggingface:xtreme/udpos.Afrikaans')

• Описание :

Universal Dependencies (UD) is a framework for consistent annotation of grammar (parts of speech, morphological
features, and syntactic dependencies) across different human languages. UD is an open community effort with over 200
contributors producing more than 100 treebanks in over 70 languages. If you’re new to UD, you should start by reading
the first part of the Short Introduction and then browsing the annotation guidelines.

The Cross-lingual TRansfer Evaluation of Multilingual Encoders (XTREME) benchmark is a benchmark for the evaluation of
the cross-lingual generalization ability of pre-trained multilingual models. It covers 40 typologically diverse languages
(spanning 12 language families) and includes nine tasks that collectively require reasoning about different levels of
syntax and semantics. The languages in XTREME are selected to maximize language diversity, coverage in existing tasks,
and availability of training data. Among these are many under-studied languages, such as the Dravidian languages Tamil
(spoken in southern India, Sri Lanka, and Singapore), Telugu and Malayalam (spoken mainly in southern India), and the
Niger-Congo languages Swahili and Yoruba, spoken in Africa.

• Лицензия : Нет известной лицензии.
• Версия : 1.0.0
• Расколы :

• Функции :

{
    "tokens": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "pos_tags": {
        "feature": {
            "num_classes": 17,
            "names": [
                "ADJ",
                "ADP",
                "ADV",
                "AUX",
                "CCONJ",
                "DET",
                "INTJ",
                "NOUN",
                "NUM",
                "PART",
                "PRON",
                "PROPN",
                "PUNCT",
                "SCONJ",
                "SYM",
                "VERB",
                "X"
            ],
            "names_file": null,
            "id": null,
            "_type": "ClassLabel"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

udpos.арабский

Используйте следующую команду, чтобы загрузить этот набор данных в TFDS:

ds = tfds.load('huggingface:xtreme/udpos.Arabic')

• Описание :

Universal Dependencies (UD) is a framework for consistent annotation of grammar (parts of speech, morphological
features, and syntactic dependencies) across different human languages. UD is an open community effort with over 200
contributors producing more than 100 treebanks in over 70 languages. If you’re new to UD, you should start by reading
the first part of the Short Introduction and then browsing the annotation guidelines.

The Cross-lingual TRansfer Evaluation of Multilingual Encoders (XTREME) benchmark is a benchmark for the evaluation of
the cross-lingual generalization ability of pre-trained multilingual models. It covers 40 typologically diverse languages
(spanning 12 language families) and includes nine tasks that collectively require reasoning about different levels of
syntax and semantics. The languages in XTREME are selected to maximize language diversity, coverage in existing tasks,
and availability of training data. Among these are many under-studied languages, such as the Dravidian languages Tamil
(spoken in southern India, Sri Lanka, and Singapore), Telugu and Malayalam (spoken mainly in southern India), and the
Niger-Congo languages Swahili and Yoruba, spoken in Africa.

• Лицензия : Нет известной лицензии.
• Версия : 1.0.0
• Расколы :

• Функции :

{
    "tokens": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "pos_tags": {
        "feature": {
            "num_classes": 17,
            "names": [
                "ADJ",
                "ADP",
                "ADV",
                "AUX",
                "CCONJ",
                "DET",
                "INTJ",
                "NOUN",
                "NUM",
                "PART",
                "PRON",
                "PROPN",
                "PUNCT",
                "SCONJ",
                "SYM",
                "VERB",
                "X"
            ],
            "names_file": null,
            "id": null,
            "_type": "ClassLabel"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

udpos.Баскский

Используйте следующую команду, чтобы загрузить этот набор данных в TFDS:

ds = tfds.load('huggingface:xtreme/udpos.Basque')

• Описание :

Universal Dependencies (UD) is a framework for consistent annotation of grammar (parts of speech, morphological
features, and syntactic dependencies) across different human languages. UD is an open community effort with over 200
contributors producing more than 100 treebanks in over 70 languages. If you’re new to UD, you should start by reading
the first part of the Short Introduction and then browsing the annotation guidelines.

The Cross-lingual TRansfer Evaluation of Multilingual Encoders (XTREME) benchmark is a benchmark for the evaluation of
the cross-lingual generalization ability of pre-trained multilingual models. It covers 40 typologically diverse languages
(spanning 12 language families) and includes nine tasks that collectively require reasoning about different levels of
syntax and semantics. The languages in XTREME are selected to maximize language diversity, coverage in existing tasks,
and availability of training data. Among these are many under-studied languages, such as the Dravidian languages Tamil
(spoken in southern India, Sri Lanka, and Singapore), Telugu and Malayalam (spoken mainly in southern India), and the
Niger-Congo languages Swahili and Yoruba, spoken in Africa.

• Лицензия : Нет известной лицензии.
• Версия : 1.0.0
• Расколы :

• Функции :

{
    "tokens": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "pos_tags": {
        "feature": {
            "num_classes": 17,
            "names": [
                "ADJ",
                "ADP",
                "ADV",
                "AUX",
                "CCONJ",
                "DET",
                "INTJ",
                "NOUN",
                "NUM",
                "PART",
                "PRON",
                "PROPN",
                "PUNCT",
                "SCONJ",
                "SYM",
                "VERB",
                "X"
            ],
            "names_file": null,
            "id": null,
            "_type": "ClassLabel"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

удпос.болгарский

Используйте следующую команду, чтобы загрузить этот набор данных в TFDS:

ds = tfds.load('huggingface:xtreme/udpos.Bulgarian')

• Описание :

Universal Dependencies (UD) is a framework for consistent annotation of grammar (parts of speech, morphological
features, and syntactic dependencies) across different human languages. UD is an open community effort with over 200
contributors producing more than 100 treebanks in over 70 languages. If you’re new to UD, you should start by reading
the first part of the Short Introduction and then browsing the annotation guidelines.

The Cross-lingual TRansfer Evaluation of Multilingual Encoders (XTREME) benchmark is a benchmark for the evaluation of
the cross-lingual generalization ability of pre-trained multilingual models. It covers 40 typologically diverse languages
(spanning 12 language families) and includes nine tasks that collectively require reasoning about different levels of
syntax and semantics. The languages in XTREME are selected to maximize language diversity, coverage in existing tasks,
and availability of training data. Among these are many under-studied languages, such as the Dravidian languages Tamil
(spoken in southern India, Sri Lanka, and Singapore), Telugu and Malayalam (spoken mainly in southern India), and the
Niger-Congo languages Swahili and Yoruba, spoken in Africa.

• Лицензия : Нет известной лицензии.
• Версия : 1.0.0
• Расколы :

• Функции :

{
    "tokens": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "pos_tags": {
        "feature": {
            "num_classes": 17,
            "names": [
                "ADJ",
                "ADP",
                "ADV",
                "AUX",
                "CCONJ",
                "DET",
                "INTJ",
                "NOUN",
                "NUM",
                "PART",
                "PRON",
                "PROPN",
                "PUNCT",
                "SCONJ",
                "SYM",
                "VERB",
                "X"
            ],
            "names_file": null,
            "id": null,
            "_type": "ClassLabel"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

udpos.голландский

Используйте следующую команду, чтобы загрузить этот набор данных в TFDS:

ds = tfds.load('huggingface:xtreme/udpos.Dutch')

• Описание :

Universal Dependencies (UD) is a framework for consistent annotation of grammar (parts of speech, morphological
features, and syntactic dependencies) across different human languages. UD is an open community effort with over 200
contributors producing more than 100 treebanks in over 70 languages. If you’re new to UD, you should start by reading
the first part of the Short Introduction and then browsing the annotation guidelines.

The Cross-lingual TRansfer Evaluation of Multilingual Encoders (XTREME) benchmark is a benchmark for the evaluation of
the cross-lingual generalization ability of pre-trained multilingual models. It covers 40 typologically diverse languages
(spanning 12 language families) and includes nine tasks that collectively require reasoning about different levels of
syntax and semantics. The languages in XTREME are selected to maximize language diversity, coverage in existing tasks,
and availability of training data. Among these are many under-studied languages, such as the Dravidian languages Tamil
(spoken in southern India, Sri Lanka, and Singapore), Telugu and Malayalam (spoken mainly in southern India), and the
Niger-Congo languages Swahili and Yoruba, spoken in Africa.

• Лицензия : Нет известной лицензии.
• Версия : 1.0.0
• Расколы :

• Функции :

{
    "tokens": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "pos_tags": {
        "feature": {
            "num_classes": 17,
            "names": [
                "ADJ",
                "ADP",
                "ADV",
                "AUX",
                "CCONJ",
                "DET",
                "INTJ",
                "NOUN",
                "NUM",
                "PART",
                "PRON",
                "PROPN",
                "PUNCT",
                "SCONJ",
                "SYM",
                "VERB",
                "X"
            ],
            "names_file": null,
            "id": null,
            "_type": "ClassLabel"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

udpos.Английский

Используйте следующую команду, чтобы загрузить этот набор данных в TFDS:

ds = tfds.load('huggingface:xtreme/udpos.English')

• Описание :

Universal Dependencies (UD) is a framework for consistent annotation of grammar (parts of speech, morphological
features, and syntactic dependencies) across different human languages. UD is an open community effort with over 200
contributors producing more than 100 treebanks in over 70 languages. If you’re new to UD, you should start by reading
the first part of the Short Introduction and then browsing the annotation guidelines.

The Cross-lingual TRansfer Evaluation of Multilingual Encoders (XTREME) benchmark is a benchmark for the evaluation of
the cross-lingual generalization ability of pre-trained multilingual models. It covers 40 typologically diverse languages
(spanning 12 language families) and includes nine tasks that collectively require reasoning about different levels of
syntax and semantics. The languages in XTREME are selected to maximize language diversity, coverage in existing tasks,
and availability of training data. Among these are many under-studied languages, such as the Dravidian languages Tamil
(spoken in southern India, Sri Lanka, and Singapore), Telugu and Malayalam (spoken mainly in southern India), and the
Niger-Congo languages Swahili and Yoruba, spoken in Africa.

• Лицензия : Нет известной лицензии.
• Версия : 1.0.0
• Расколы :

• Функции :

{
    "tokens": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "pos_tags": {
        "feature": {
            "num_classes": 17,
            "names": [
                "ADJ",
                "ADP",
                "ADV",
                "AUX",
                "CCONJ",
                "DET",
                "INTJ",
                "NOUN",
                "NUM",
                "PART",
                "PRON",
                "PROPN",
                "PUNCT",
                "SCONJ",
                "SYM",
                "VERB",
                "X"
            ],
            "names_file": null,
            "id": null,
            "_type": "ClassLabel"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

udpos.эстонский

Используйте следующую команду, чтобы загрузить этот набор данных в TFDS:

ds = tfds.load('huggingface:xtreme/udpos.Estonian')

• Описание :

Universal Dependencies (UD) is a framework for consistent annotation of grammar (parts of speech, morphological
features, and syntactic dependencies) across different human languages. UD is an open community effort with over 200
contributors producing more than 100 treebanks in over 70 languages. If you’re new to UD, you should start by reading
the first part of the Short Introduction and then browsing the annotation guidelines.

The Cross-lingual TRansfer Evaluation of Multilingual Encoders (XTREME) benchmark is a benchmark for the evaluation of
the cross-lingual generalization ability of pre-trained multilingual models. It covers 40 typologically diverse languages
(spanning 12 language families) and includes nine tasks that collectively require reasoning about different levels of
syntax and semantics. The languages in XTREME are selected to maximize language diversity, coverage in existing tasks,
and availability of training data. Among these are many under-studied languages, such as the Dravidian languages Tamil
(spoken in southern India, Sri Lanka, and Singapore), Telugu and Malayalam (spoken mainly in southern India), and the
Niger-Congo languages Swahili and Yoruba, spoken in Africa.

• Лицензия : Нет известной лицензии.
• Версия : 1.0.0
• Расколы :

• Функции :

{
    "tokens": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "pos_tags": {
        "feature": {
            "num_classes": 17,
            "names": [
                "ADJ",
                "ADP",
                "ADV",
                "AUX",
                "CCONJ",
                "DET",
                "INTJ",
                "NOUN",
                "NUM",
                "PART",
                "PRON",
                "PROPN",
                "PUNCT",
                "SCONJ",
                "SYM",
                "VERB",
                "X"
            ],
            "names_file": null,
            "id": null,
            "_type": "ClassLabel"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

udpos.финский

Используйте следующую команду, чтобы загрузить этот набор данных в TFDS:

ds = tfds.load('huggingface:xtreme/udpos.Finnish')

• Описание :

Universal Dependencies (UD) is a framework for consistent annotation of grammar (parts of speech, morphological
features, and syntactic dependencies) across different human languages. UD is an open community effort with over 200
contributors producing more than 100 treebanks in over 70 languages. If you’re new to UD, you should start by reading
the first part of the Short Introduction and then browsing the annotation guidelines.

The Cross-lingual TRansfer Evaluation of Multilingual Encoders (XTREME) benchmark is a benchmark for the evaluation of
the cross-lingual generalization ability of pre-trained multilingual models. It covers 40 typologically diverse languages
(spanning 12 language families) and includes nine tasks that collectively require reasoning about different levels of
syntax and semantics. The languages in XTREME are selected to maximize language diversity, coverage in existing tasks,
and availability of training data. Among these are many under-studied languages, such as the Dravidian languages Tamil
(spoken in southern India, Sri Lanka, and Singapore), Telugu and Malayalam (spoken mainly in southern India), and the
Niger-Congo languages Swahili and Yoruba, spoken in Africa.

• Лицензия : Нет известной лицензии.
• Версия : 1.0.0
• Расколы :

• Функции :

{
    "tokens": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "pos_tags": {
        "feature": {
            "num_classes": 17,
            "names": [
                "ADJ",
                "ADP",
                "ADV",
                "AUX",
                "CCONJ",
                "DET",
                "INTJ",
                "NOUN",
                "NUM",
                "PART",
                "PRON",
                "PROPN",
                "PUNCT",
                "SCONJ",
                "SYM",
                "VERB",
                "X"
            ],
            "names_file": null,
            "id": null,
            "_type": "ClassLabel"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

udpos.Французский

Используйте следующую команду, чтобы загрузить этот набор данных в TFDS:

ds = tfds.load('huggingface:xtreme/udpos.French')

• Описание :

Universal Dependencies (UD) is a framework for consistent annotation of grammar (parts of speech, morphological
features, and syntactic dependencies) across different human languages. UD is an open community effort with over 200
contributors producing more than 100 treebanks in over 70 languages. If you’re new to UD, you should start by reading
the first part of the Short Introduction and then browsing the annotation guidelines.

The Cross-lingual TRansfer Evaluation of Multilingual Encoders (XTREME) benchmark is a benchmark for the evaluation of
the cross-lingual generalization ability of pre-trained multilingual models. It covers 40 typologically diverse languages
(spanning 12 language families) and includes nine tasks that collectively require reasoning about different levels of
syntax and semantics. The languages in XTREME are selected to maximize language diversity, coverage in existing tasks,
and availability of training data. Among these are many under-studied languages, such as the Dravidian languages Tamil
(spoken in southern India, Sri Lanka, and Singapore), Telugu and Malayalam (spoken mainly in southern India), and the
Niger-Congo languages Swahili and Yoruba, spoken in Africa.

• Лицензия : Нет известной лицензии.
• Версия : 1.0.0
• Расколы :

• Функции :

{
    "tokens": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "pos_tags": {
        "feature": {
            "num_classes": 17,
            "names": [
                "ADJ",
                "ADP",
                "ADV",
                "AUX",
                "CCONJ",
                "DET",
                "INTJ",
                "NOUN",
                "NUM",
                "PART",
                "PRON",
                "PROPN",
                "PUNCT",
                "SCONJ",
                "SYM",
                "VERB",
                "X"
            ],
            "names_file": null,
            "id": null,
            "_type": "ClassLabel"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

udpos.немецкий

Используйте следующую команду, чтобы загрузить этот набор данных в TFDS:

ds = tfds.load('huggingface:xtreme/udpos.German')

• Описание :

Universal Dependencies (UD) is a framework for consistent annotation of grammar (parts of speech, morphological
features, and syntactic dependencies) across different human languages. UD is an open community effort with over 200
contributors producing more than 100 treebanks in over 70 languages. If you’re new to UD, you should start by reading
the first part of the Short Introduction and then browsing the annotation guidelines.

The Cross-lingual TRansfer Evaluation of Multilingual Encoders (XTREME) benchmark is a benchmark for the evaluation of
the cross-lingual generalization ability of pre-trained multilingual models. It covers 40 typologically diverse languages
(spanning 12 language families) and includes nine tasks that collectively require reasoning about different levels of
syntax and semantics. The languages in XTREME are selected to maximize language diversity, coverage in existing tasks,
and availability of training data. Among these are many under-studied languages, such as the Dravidian languages Tamil
(spoken in southern India, Sri Lanka, and Singapore), Telugu and Malayalam (spoken mainly in southern India), and the
Niger-Congo languages Swahili and Yoruba, spoken in Africa.

• Лицензия : Нет известной лицензии.
• Версия : 1.0.0
• Расколы :

• Функции :

{
    "tokens": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "pos_tags": {
        "feature": {
            "num_classes": 17,
            "names": [
                "ADJ",
                "ADP",
                "ADV",
                "AUX",
                "CCONJ",
                "DET",
                "INTJ",
                "NOUN",
                "NUM",
                "PART",
                "PRON",
                "PROPN",
                "PUNCT",
                "SCONJ",
                "SYM",
                "VERB",
                "X"
            ],
            "names_file": null,
            "id": null,
            "_type": "ClassLabel"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

udpos.греческий

Используйте следующую команду, чтобы загрузить этот набор данных в TFDS:

ds = tfds.load('huggingface:xtreme/udpos.Greek')

• Описание :

Universal Dependencies (UD) is a framework for consistent annotation of grammar (parts of speech, morphological
features, and syntactic dependencies) across different human languages. UD is an open community effort with over 200
contributors producing more than 100 treebanks in over 70 languages. If you’re new to UD, you should start by reading
the first part of the Short Introduction and then browsing the annotation guidelines.

The Cross-lingual TRansfer Evaluation of Multilingual Encoders (XTREME) benchmark is a benchmark for the evaluation of
the cross-lingual generalization ability of pre-trained multilingual models. It covers 40 typologically diverse languages
(spanning 12 language families) and includes nine tasks that collectively require reasoning about different levels of
syntax and semantics. The languages in XTREME are selected to maximize language diversity, coverage in existing tasks,
and availability of training data. Among these are many under-studied languages, such as the Dravidian languages Tamil
(spoken in southern India, Sri Lanka, and Singapore), Telugu and Malayalam (spoken mainly in southern India), and the
Niger-Congo languages Swahili and Yoruba, spoken in Africa.

• Лицензия : Нет известной лицензии.
• Версия : 1.0.0
• Расколы :

• Функции :

{
    "tokens": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "pos_tags": {
        "feature": {
            "num_classes": 17,
            "names": [
                "ADJ",
                "ADP",
                "ADV",
                "AUX",
                "CCONJ",
                "DET",
                "INTJ",
                "NOUN",
                "NUM",
                "PART",
                "PRON",
                "PROPN",
                "PUNCT",
                "SCONJ",
                "SYM",
                "VERB",
                "X"
            ],
            "names_file": null,
            "id": null,
            "_type": "ClassLabel"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

udpos.иврит

Используйте следующую команду, чтобы загрузить этот набор данных в TFDS:

ds = tfds.load('huggingface:xtreme/udpos.Hebrew')

• Описание :

Universal Dependencies (UD) is a framework for consistent annotation of grammar (parts of speech, morphological
features, and syntactic dependencies) across different human languages. UD is an open community effort with over 200
contributors producing more than 100 treebanks in over 70 languages. If you’re new to UD, you should start by reading
the first part of the Short Introduction and then browsing the annotation guidelines.

The Cross-lingual TRansfer Evaluation of Multilingual Encoders (XTREME) benchmark is a benchmark for the evaluation of
the cross-lingual generalization ability of pre-trained multilingual models. It covers 40 typologically diverse languages
(spanning 12 language families) and includes nine tasks that collectively require reasoning about different levels of
syntax and semantics. The languages in XTREME are selected to maximize language diversity, coverage in existing tasks,
and availability of training data. Among these are many under-studied languages, such as the Dravidian languages Tamil
(spoken in southern India, Sri Lanka, and Singapore), Telugu and Malayalam (spoken mainly in southern India), and the
Niger-Congo languages Swahili and Yoruba, spoken in Africa.

• Лицензия : Нет известной лицензии.
• Версия : 1.0.0
• Расколы :

• Функции :

{
    "tokens": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "pos_tags": {
        "feature": {
            "num_classes": 17,
            "names": [
                "ADJ",
                "ADP",
                "ADV",
                "AUX",
                "CCONJ",
                "DET",
                "INTJ",
                "NOUN",
                "NUM",
                "PART",
                "PRON",
                "PROPN",
                "PUNCT",
                "SCONJ",
                "SYM",
                "VERB",
                "X"
            ],
            "names_file": null,
            "id": null,
            "_type": "ClassLabel"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

udpos.хинди

Используйте следующую команду, чтобы загрузить этот набор данных в TFDS:

ds = tfds.load('huggingface:xtreme/udpos.Hindi')

• Описание :

Universal Dependencies (UD) is a framework for consistent annotation of grammar (parts of speech, morphological
features, and syntactic dependencies) across different human languages. UD is an open community effort with over 200
contributors producing more than 100 treebanks in over 70 languages. If you’re new to UD, you should start by reading
the first part of the Short Introduction and then browsing the annotation guidelines.

The Cross-lingual TRansfer Evaluation of Multilingual Encoders (XTREME) benchmark is a benchmark for the evaluation of
the cross-lingual generalization ability of pre-trained multilingual models. It covers 40 typologically diverse languages
(spanning 12 language families) and includes nine tasks that collectively require reasoning about different levels of
syntax and semantics. The languages in XTREME are selected to maximize language diversity, coverage in existing tasks,
and availability of training data. Among these are many under-studied languages, such as the Dravidian languages Tamil
(spoken in southern India, Sri Lanka, and Singapore), Telugu and Malayalam (spoken mainly in southern India), and the
Niger-Congo languages Swahili and Yoruba, spoken in Africa.

• Лицензия : Нет известной лицензии.
• Версия : 1.0.0
• Расколы :

• Функции :

{
    "tokens": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "pos_tags": {
        "feature": {
            "num_classes": 17,
            "names": [
                "ADJ",
                "ADP",
                "ADV",
                "AUX",
                "CCONJ",
                "DET",
                "INTJ",
                "NOUN",
                "NUM",
                "PART",
                "PRON",
                "PROPN",
                "PUNCT",
                "SCONJ",
                "SYM",
                "VERB",
                "X"
            ],
            "names_file": null,
            "id": null,
            "_type": "ClassLabel"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

udpos.Венгерский

Используйте следующую команду, чтобы загрузить этот набор данных в TFDS:

ds = tfds.load('huggingface:xtreme/udpos.Hungarian')

• Описание :

Universal Dependencies (UD) is a framework for consistent annotation of grammar (parts of speech, morphological
features, and syntactic dependencies) across different human languages. UD is an open community effort with over 200
contributors producing more than 100 treebanks in over 70 languages. If you’re new to UD, you should start by reading
the first part of the Short Introduction and then browsing the annotation guidelines.

The Cross-lingual TRansfer Evaluation of Multilingual Encoders (XTREME) benchmark is a benchmark for the evaluation of
the cross-lingual generalization ability of pre-trained multilingual models. It covers 40 typologically diverse languages
(spanning 12 language families) and includes nine tasks that collectively require reasoning about different levels of
syntax and semantics. The languages in XTREME are selected to maximize language diversity, coverage in existing tasks,
and availability of training data. Among these are many under-studied languages, such as the Dravidian languages Tamil
(spoken in southern India, Sri Lanka, and Singapore), Telugu and Malayalam (spoken mainly in southern India), and the
Niger-Congo languages Swahili and Yoruba, spoken in Africa.

• Лицензия : Нет известной лицензии.
• Версия : 1.0.0
• Расколы :

• Функции :

{
    "tokens": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "pos_tags": {
        "feature": {
            "num_classes": 17,
            "names": [
                "ADJ",
                "ADP",
                "ADV",
                "AUX",
                "CCONJ",
                "DET",
                "INTJ",
                "NOUN",
                "NUM",
                "PART",
                "PRON",
                "PROPN",
                "PUNCT",
                "SCONJ",
                "SYM",
                "VERB",
                "X"
            ],
            "names_file": null,
            "id": null,
            "_type": "ClassLabel"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

udpos.индонезийский

Используйте следующую команду, чтобы загрузить этот набор данных в TFDS:

ds = tfds.load('huggingface:xtreme/udpos.Indonesian')

• Описание :

Universal Dependencies (UD) is a framework for consistent annotation of grammar (parts of speech, morphological
features, and syntactic dependencies) across different human languages. UD is an open community effort with over 200
contributors producing more than 100 treebanks in over 70 languages. If you’re new to UD, you should start by reading
the first part of the Short Introduction and then browsing the annotation guidelines.

The Cross-lingual TRansfer Evaluation of Multilingual Encoders (XTREME) benchmark is a benchmark for the evaluation of
the cross-lingual generalization ability of pre-trained multilingual models. It covers 40 typologically diverse languages
(spanning 12 language families) and includes nine tasks that collectively require reasoning about different levels of
syntax and semantics. The languages in XTREME are selected to maximize language diversity, coverage in existing tasks,
and availability of training data. Among these are many under-studied languages, such as the Dravidian languages Tamil
(spoken in southern India, Sri Lanka, and Singapore), Telugu and Malayalam (spoken mainly in southern India), and the
Niger-Congo languages Swahili and Yoruba, spoken in Africa.

• Лицензия : Нет известной лицензии.
• Версия : 1.0.0
• Расколы :

• Функции :

{
    "tokens": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "pos_tags": {
        "feature": {
            "num_classes": 17,
            "names": [
                "ADJ",
                "ADP",
                "ADV",
                "AUX",
                "CCONJ",
                "DET",
                "INTJ",
                "NOUN",
                "NUM",
                "PART",
                "PRON",
                "PROPN",
                "PUNCT",
                "SCONJ",
                "SYM",
                "VERB",
                "X"
            ],
            "names_file": null,
            "id": null,
            "_type": "ClassLabel"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

udpos.итальянский

Используйте следующую команду, чтобы загрузить этот набор данных в TFDS:

ds = tfds.load('huggingface:xtreme/udpos.Italian')

• Описание :

Universal Dependencies (UD) is a framework for consistent annotation of grammar (parts of speech, morphological
features, and syntactic dependencies) across different human languages. UD is an open community effort with over 200
contributors producing more than 100 treebanks in over 70 languages. If you’re new to UD, you should start by reading
the first part of the Short Introduction and then browsing the annotation guidelines.

The Cross-lingual TRansfer Evaluation of Multilingual Encoders (XTREME) benchmark is a benchmark for the evaluation of
the cross-lingual generalization ability of pre-trained multilingual models. It covers 40 typologically diverse languages
(spanning 12 language families) and includes nine tasks that collectively require reasoning about different levels of
syntax and semantics. The languages in XTREME are selected to maximize language diversity, coverage in existing tasks,
and availability of training data. Among these are many under-studied languages, such as the Dravidian languages Tamil
(spoken in southern India, Sri Lanka, and Singapore), Telugu and Malayalam (spoken mainly in southern India), and the
Niger-Congo languages Swahili and Yoruba, spoken in Africa.

• Лицензия : Нет известной лицензии.
• Версия : 1.0.0
• Расколы :

• Функции :

{
    "tokens": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "pos_tags": {
        "feature": {
            "num_classes": 17,
            "names": [
                "ADJ",
                "ADP",
                "ADV",
                "AUX",
                "CCONJ",
                "DET",
                "INTJ",
                "NOUN",
                "NUM",
                "PART",
                "PRON",
                "PROPN",
                "PUNCT",
                "SCONJ",
                "SYM",
                "VERB",
                "X"
            ],
            "names_file": null,
            "id": null,
            "_type": "ClassLabel"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

udpos.Японский

Используйте следующую команду, чтобы загрузить этот набор данных в TFDS:

ds = tfds.load('huggingface:xtreme/udpos.Japanese')

• Описание :

Universal Dependencies (UD) is a framework for consistent annotation of grammar (parts of speech, morphological
features, and syntactic dependencies) across different human languages. UD is an open community effort with over 200
contributors producing more than 100 treebanks in over 70 languages. If you’re new to UD, you should start by reading
the first part of the Short Introduction and then browsing the annotation guidelines.

The Cross-lingual TRansfer Evaluation of Multilingual Encoders (XTREME) benchmark is a benchmark for the evaluation of
the cross-lingual generalization ability of pre-trained multilingual models. It covers 40 typologically diverse languages
(spanning 12 language families) and includes nine tasks that collectively require reasoning about different levels of
syntax and semantics. The languages in XTREME are selected to maximize language diversity, coverage in existing tasks,
and availability of training data. Among these are many under-studied languages, such as the Dravidian languages Tamil
(spoken in southern India, Sri Lanka, and Singapore), Telugu and Malayalam (spoken mainly in southern India), and the
Niger-Congo languages Swahili and Yoruba, spoken in Africa.

• Лицензия : Нет известной лицензии.
• Версия : 1.0.0
• Расколы :

• Функции :

{
    "tokens": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "pos_tags": {
        "feature": {
            "num_classes": 17,
            "names": [
                "ADJ",
                "ADP",
                "ADV",
                "AUX",
                "CCONJ",
                "DET",
                "INTJ",
                "NOUN",
                "NUM",
                "PART",
                "PRON",
                "PROPN",
                "PUNCT",
                "SCONJ",
                "SYM",
                "VERB",
                "X"
            ],
            "names_file": null,
            "id": null,
            "_type": "ClassLabel"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

удпос.Казахский

Используйте следующую команду, чтобы загрузить этот набор данных в TFDS:

ds = tfds.load('huggingface:xtreme/udpos.Kazakh')

• Описание :

Universal Dependencies (UD) is a framework for consistent annotation of grammar (parts of speech, morphological
features, and syntactic dependencies) across different human languages. UD is an open community effort with over 200
contributors producing more than 100 treebanks in over 70 languages. If you’re new to UD, you should start by reading
the first part of the Short Introduction and then browsing the annotation guidelines.

The Cross-lingual TRansfer Evaluation of Multilingual Encoders (XTREME) benchmark is a benchmark for the evaluation of
the cross-lingual generalization ability of pre-trained multilingual models. It covers 40 typologically diverse languages
(spanning 12 language families) and includes nine tasks that collectively require reasoning about different levels of
syntax and semantics. The languages in XTREME are selected to maximize language diversity, coverage in existing tasks,
and availability of training data. Among these are many under-studied languages, such as the Dravidian languages Tamil
(spoken in southern India, Sri Lanka, and Singapore), Telugu and Malayalam (spoken mainly in southern India), and the
Niger-Congo languages Swahili and Yoruba, spoken in Africa.

• Лицензия : Нет известной лицензии.
• Версия : 1.0.0
• Расколы :

• Функции :

{
    "tokens": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "pos_tags": {
        "feature": {
            "num_classes": 17,
            "names": [
                "ADJ",
                "ADP",
                "ADV",
                "AUX",
                "CCONJ",
                "DET",
                "INTJ",
                "NOUN",
                "NUM",
                "PART",
                "PRON",
                "PROPN",
                "PUNCT",
                "SCONJ",
                "SYM",
                "VERB",
                "X"
            ],
            "names_file": null,
            "id": null,
            "_type": "ClassLabel"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

udpos.корейский

Используйте следующую команду, чтобы загрузить этот набор данных в TFDS:

ds = tfds.load('huggingface:xtreme/udpos.Korean')

• Описание :

Universal Dependencies (UD) is a framework for consistent annotation of grammar (parts of speech, morphological
features, and syntactic dependencies) across different human languages. UD is an open community effort with over 200
contributors producing more than 100 treebanks in over 70 languages. If you’re new to UD, you should start by reading
the first part of the Short Introduction and then browsing the annotation guidelines.

The Cross-lingual TRansfer Evaluation of Multilingual Encoders (XTREME) benchmark is a benchmark for the evaluation of
the cross-lingual generalization ability of pre-trained multilingual models. It covers 40 typologically diverse languages
(spanning 12 language families) and includes nine tasks that collectively require reasoning about different levels of
syntax and semantics. The languages in XTREME are selected to maximize language diversity, coverage in existing tasks,
and availability of training data. Among these are many under-studied languages, such as the Dravidian languages Tamil
(spoken in southern India, Sri Lanka, and Singapore), Telugu and Malayalam (spoken mainly in southern India), and the
Niger-Congo languages Swahili and Yoruba, spoken in Africa.

• Лицензия : Нет известной лицензии.
• Версия : 1.0.0
• Расколы :

• Функции :

{
    "tokens": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "pos_tags": {
        "feature": {
            "num_classes": 17,
            "names": [
                "ADJ",
                "ADP",
                "ADV",
                "AUX",
                "CCONJ",
                "DET",
                "INTJ",
                "NOUN",
                "NUM",
                "PART",
                "PRON",
                "PROPN",
                "PUNCT",
                "SCONJ",
                "SYM",
                "VERB",
                "X"
            ],
            "names_file": null,
            "id": null,
            "_type": "ClassLabel"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

udpos.Китайский

Используйте следующую команду, чтобы загрузить этот набор данных в TFDS:

ds = tfds.load('huggingface:xtreme/udpos.Chinese')

• Описание :

Universal Dependencies (UD) is a framework for consistent annotation of grammar (parts of speech, morphological
features, and syntactic dependencies) across different human languages. UD is an open community effort with over 200
contributors producing more than 100 treebanks in over 70 languages. If you’re new to UD, you should start by reading
the first part of the Short Introduction and then browsing the annotation guidelines.

The Cross-lingual TRansfer Evaluation of Multilingual Encoders (XTREME) benchmark is a benchmark for the evaluation of
the cross-lingual generalization ability of pre-trained multilingual models. It covers 40 typologically diverse languages
(spanning 12 language families) and includes nine tasks that collectively require reasoning about different levels of
syntax and semantics. The languages in XTREME are selected to maximize language diversity, coverage in existing tasks,
and availability of training data. Among these are many under-studied languages, such as the Dravidian languages Tamil
(spoken in southern India, Sri Lanka, and Singapore), Telugu and Malayalam (spoken mainly in southern India), and the
Niger-Congo languages Swahili and Yoruba, spoken in Africa.

• Лицензия : Нет известной лицензии.
• Версия : 1.0.0
• Расколы :

• Функции :

{
    "tokens": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "pos_tags": {
        "feature": {
            "num_classes": 17,
            "names": [
                "ADJ",
                "ADP",
                "ADV",
                "AUX",
                "CCONJ",
                "DET",
                "INTJ",
                "NOUN",
                "NUM",
                "PART",
                "PRON",
                "PROPN",
                "PUNCT",
                "SCONJ",
                "SYM",
                "VERB",
                "X"
            ],
            "names_file": null,
            "id": null,
            "_type": "ClassLabel"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

удпос.маратхи

Используйте следующую команду, чтобы загрузить этот набор данных в TFDS:

ds = tfds.load('huggingface:xtreme/udpos.Marathi')

• Описание :

Universal Dependencies (UD) is a framework for consistent annotation of grammar (parts of speech, morphological
features, and syntactic dependencies) across different human languages. UD is an open community effort with over 200
contributors producing more than 100 treebanks in over 70 languages. If you’re new to UD, you should start by reading
the first part of the Short Introduction and then browsing the annotation guidelines.

The Cross-lingual TRansfer Evaluation of Multilingual Encoders (XTREME) benchmark is a benchmark for the evaluation of
the cross-lingual generalization ability of pre-trained multilingual models. It covers 40 typologically diverse languages
(spanning 12 language families) and includes nine tasks that collectively require reasoning about different levels of
syntax and semantics. The languages in XTREME are selected to maximize language diversity, coverage in existing tasks,
and availability of training data. Among these are many under-studied languages, such as the Dravidian languages Tamil
(spoken in southern India, Sri Lanka, and Singapore), Telugu and Malayalam (spoken mainly in southern India), and the
Niger-Congo languages Swahili and Yoruba, spoken in Africa.

• Лицензия : Нет известной лицензии.
• Версия : 1.0.0
• Расколы :

• Функции :

{
    "tokens": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "pos_tags": {
        "feature": {
            "num_classes": 17,
            "names": [
                "ADJ",
                "ADP",
                "ADV",
                "AUX",
                "CCONJ",
                "DET",
                "INTJ",
                "NOUN",
                "NUM",
                "PART",
                "PRON",
                "PROPN",
                "PUNCT",
                "SCONJ",
                "SYM",
                "VERB",
                "X"
            ],
            "names_file": null,
            "id": null,
            "_type": "ClassLabel"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

udpos.персидский

Используйте следующую команду, чтобы загрузить этот набор данных в TFDS:

ds = tfds.load('huggingface:xtreme/udpos.Persian')

• Описание :

Universal Dependencies (UD) is a framework for consistent annotation of grammar (parts of speech, morphological
features, and syntactic dependencies) across different human languages. UD is an open community effort with over 200
contributors producing more than 100 treebanks in over 70 languages. If you’re new to UD, you should start by reading
the first part of the Short Introduction and then browsing the annotation guidelines.

The Cross-lingual TRansfer Evaluation of Multilingual Encoders (XTREME) benchmark is a benchmark for the evaluation of
the cross-lingual generalization ability of pre-trained multilingual models. It covers 40 typologically diverse languages
(spanning 12 language families) and includes nine tasks that collectively require reasoning about different levels of
syntax and semantics. The languages in XTREME are selected to maximize language diversity, coverage in existing tasks,
and availability of training data. Among these are many under-studied languages, such as the Dravidian languages Tamil
(spoken in southern India, Sri Lanka, and Singapore), Telugu and Malayalam (spoken mainly in southern India), and the
Niger-Congo languages Swahili and Yoruba, spoken in Africa.

• Лицензия : Нет известной лицензии.
• Версия : 1.0.0
• Расколы :

• Функции :

{
    "tokens": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "pos_tags": {
        "feature": {
            "num_classes": 17,
            "names": [
                "ADJ",
                "ADP",
                "ADV",
                "AUX",
                "CCONJ",
                "DET",
                "INTJ",
                "NOUN",
                "NUM",
                "PART",
                "PRON",
                "PROPN",
                "PUNCT",
                "SCONJ",
                "SYM",
                "VERB",
                "X"
            ],
            "names_file": null,
            "id": null,
            "_type": "ClassLabel"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

udpos.Португальский

Используйте следующую команду, чтобы загрузить этот набор данных в TFDS:

ds = tfds.load('huggingface:xtreme/udpos.Portuguese')

• Описание :

Universal Dependencies (UD) is a framework for consistent annotation of grammar (parts of speech, morphological
features, and syntactic dependencies) across different human languages. UD is an open community effort with over 200
contributors producing more than 100 treebanks in over 70 languages. If you’re new to UD, you should start by reading
the first part of the Short Introduction and then browsing the annotation guidelines.

The Cross-lingual TRansfer Evaluation of Multilingual Encoders (XTREME) benchmark is a benchmark for the evaluation of
the cross-lingual generalization ability of pre-trained multilingual models. It covers 40 typologically diverse languages
(spanning 12 language families) and includes nine tasks that collectively require reasoning about different levels of
syntax and semantics. The languages in XTREME are selected to maximize language diversity, coverage in existing tasks,
and availability of training data. Among these are many under-studied languages, such as the Dravidian languages Tamil
(spoken in southern India, Sri Lanka, and Singapore), Telugu and Malayalam (spoken mainly in southern India), and the
Niger-Congo languages Swahili and Yoruba, spoken in Africa.

• Лицензия : Нет известной лицензии.
• Версия : 1.0.0
• Расколы :

• Функции :

{
    "tokens": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "pos_tags": {
        "feature": {
            "num_classes": 17,
            "names": [
                "ADJ",
                "ADP",
                "ADV",
                "AUX",
                "CCONJ",
                "DET",
                "INTJ",
                "NOUN",
                "NUM",
                "PART",
                "PRON",
                "PROPN",
                "PUNCT",
                "SCONJ",
                "SYM",
                "VERB",
                "X"
            ],
            "names_file": null,
            "id": null,
            "_type": "ClassLabel"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

удпос.Русский

Используйте следующую команду, чтобы загрузить этот набор данных в TFDS:

ds = tfds.load('huggingface:xtreme/udpos.Russian')

• Описание :

Universal Dependencies (UD) is a framework for consistent annotation of grammar (parts of speech, morphological
features, and syntactic dependencies) across different human languages. UD is an open community effort with over 200
contributors producing more than 100 treebanks in over 70 languages. If you’re new to UD, you should start by reading
the first part of the Short Introduction and then browsing the annotation guidelines.

The Cross-lingual TRansfer Evaluation of Multilingual Encoders (XTREME) benchmark is a benchmark for the evaluation of
the cross-lingual generalization ability of pre-trained multilingual models. It covers 40 typologically diverse languages
(spanning 12 language families) and includes nine tasks that collectively require reasoning about different levels of
syntax and semantics. The languages in XTREME are selected to maximize language diversity, coverage in existing tasks,
and availability of training data. Among these are many under-studied languages, such as the Dravidian languages Tamil
(spoken in southern India, Sri Lanka, and Singapore), Telugu and Malayalam (spoken mainly in southern India), and the
Niger-Congo languages Swahili and Yoruba, spoken in Africa.

• Лицензия : Нет известной лицензии.
• Версия : 1.0.0
• Расколы :

• Функции :

{
    "tokens": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "pos_tags": {
        "feature": {
            "num_classes": 17,
            "names": [
                "ADJ",
                "ADP",
                "ADV",
                "AUX",
                "CCONJ",
                "DET",
                "INTJ",
                "NOUN",
                "NUM",
                "PART",
                "PRON",
                "PROPN",
                "PUNCT",
                "SCONJ",
                "SYM",
                "VERB",
                "X"
            ],
            "names_file": null,
            "id": null,
            "_type": "ClassLabel"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

udpos.Испанский

Используйте следующую команду, чтобы загрузить этот набор данных в TFDS:

ds = tfds.load('huggingface:xtreme/udpos.Spanish')

• Описание :

Universal Dependencies (UD) is a framework for consistent annotation of grammar (parts of speech, morphological
features, and syntactic dependencies) across different human languages. UD is an open community effort with over 200
contributors producing more than 100 treebanks in over 70 languages. If you’re new to UD, you should start by reading
the first part of the Short Introduction and then browsing the annotation guidelines.

The Cross-lingual TRansfer Evaluation of Multilingual Encoders (XTREME) benchmark is a benchmark for the evaluation of
the cross-lingual generalization ability of pre-trained multilingual models. It covers 40 typologically diverse languages
(spanning 12 language families) and includes nine tasks that collectively require reasoning about different levels of
syntax and semantics. The languages in XTREME are selected to maximize language diversity, coverage in existing tasks,
and availability of training data. Among these are many under-studied languages, such as the Dravidian languages Tamil
(spoken in southern India, Sri Lanka, and Singapore), Telugu and Malayalam (spoken mainly in southern India), and the
Niger-Congo languages Swahili and Yoruba, spoken in Africa.

• Лицензия : Нет известной лицензии.
• Версия : 1.0.0
• Расколы :

• Функции :

{
    "tokens": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "pos_tags": {
        "feature": {
            "num_classes": 17,
            "names": [
                "ADJ",
                "ADP",
                "ADV",
                "AUX",
                "CCONJ",
                "DET",
                "INTJ",
                "NOUN",
                "NUM",
                "PART",
                "PRON",
                "PROPN",
                "PUNCT",
                "SCONJ",
                "SYM",
                "VERB",
                "X"
            ],
            "names_file": null,
            "id": null,
            "_type": "ClassLabel"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

udpos.Тагальский

Используйте следующую команду, чтобы загрузить этот набор данных в TFDS:

ds = tfds.load('huggingface:xtreme/udpos.Tagalog')

• Описание :

Universal Dependencies (UD) is a framework for consistent annotation of grammar (parts of speech, morphological
features, and syntactic dependencies) across different human languages. UD is an open community effort with over 200
contributors producing more than 100 treebanks in over 70 languages. If you’re new to UD, you should start by reading
the first part of the Short Introduction and then browsing the annotation guidelines.

The Cross-lingual TRansfer Evaluation of Multilingual Encoders (XTREME) benchmark is a benchmark for the evaluation of
the cross-lingual generalization ability of pre-trained multilingual models. It covers 40 typologically diverse languages
(spanning 12 language families) and includes nine tasks that collectively require reasoning about different levels of
syntax and semantics. The languages in XTREME are selected to maximize language diversity, coverage in existing tasks,
and availability of training data. Among these are many under-studied languages, such as the Dravidian languages Tamil
(spoken in southern India, Sri Lanka, and Singapore), Telugu and Malayalam (spoken mainly in southern India), and the
Niger-Congo languages Swahili and Yoruba, spoken in Africa.

• Лицензия : Нет известной лицензии.
• Версия : 1.0.0
• Расколы :

• Функции :

{
    "tokens": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "pos_tags": {
        "feature": {
            "num_classes": 17,
            "names": [
                "ADJ",
                "ADP",
                "ADV",
                "AUX",
                "CCONJ",
                "DET",
                "INTJ",
                "NOUN",
                "NUM",
                "PART",
                "PRON",
                "PROPN",
                "PUNCT",
                "SCONJ",
                "SYM",
                "VERB",
                "X"
            ],
            "names_file": null,
            "id": null,
            "_type": "ClassLabel"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

udpos.Тамильский

Используйте следующую команду, чтобы загрузить этот набор данных в TFDS:

ds = tfds.load('huggingface:xtreme/udpos.Tamil')

• Описание :

Universal Dependencies (UD) is a framework for consistent annotation of grammar (parts of speech, morphological
features, and syntactic dependencies) across different human languages. UD is an open community effort with over 200
contributors producing more than 100 treebanks in over 70 languages. If you’re new to UD, you should start by reading
the first part of the Short Introduction and then browsing the annotation guidelines.

The Cross-lingual TRansfer Evaluation of Multilingual Encoders (XTREME) benchmark is a benchmark for the evaluation of
the cross-lingual generalization ability of pre-trained multilingual models. It covers 40 typologically diverse languages
(spanning 12 language families) and includes nine tasks that collectively require reasoning about different levels of
syntax and semantics. The languages in XTREME are selected to maximize language diversity, coverage in existing tasks,
and availability of training data. Among these are many under-studied languages, such as the Dravidian languages Tamil
(spoken in southern India, Sri Lanka, and Singapore), Telugu and Malayalam (spoken mainly in southern India), and the
Niger-Congo languages Swahili and Yoruba, spoken in Africa.

• Лицензия : Нет известной лицензии.
• Версия : 1.0.0
• Расколы :

• Функции :

{
    "tokens": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "pos_tags": {
        "feature": {
            "num_classes": 17,
            "names": [
                "ADJ",
                "ADP",
                "ADV",
                "AUX",
                "CCONJ",
                "DET",
                "INTJ",
                "NOUN",
                "NUM",
                "PART",
                "PRON",
                "PROPN",
                "PUNCT",
                "SCONJ",
                "SYM",
                "VERB",
                "X"
            ],
            "names_file": null,
            "id": null,
            "_type": "ClassLabel"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

udpos.телугу

Используйте следующую команду, чтобы загрузить этот набор данных в TFDS:

ds = tfds.load('huggingface:xtreme/udpos.Telugu')

• Описание :

Universal Dependencies (UD) is a framework for consistent annotation of grammar (parts of speech, morphological
features, and syntactic dependencies) across different human languages. UD is an open community effort with over 200
contributors producing more than 100 treebanks in over 70 languages. If you’re new to UD, you should start by reading
the first part of the Short Introduction and then browsing the annotation guidelines.

The Cross-lingual TRansfer Evaluation of Multilingual Encoders (XTREME) benchmark is a benchmark for the evaluation of
the cross-lingual generalization ability of pre-trained multilingual models. It covers 40 typologically diverse languages
(spanning 12 language families) and includes nine tasks that collectively require reasoning about different levels of
syntax and semantics. The languages in XTREME are selected to maximize language diversity, coverage in existing tasks,
and availability of training data. Among these are many under-studied languages, such as the Dravidian languages Tamil
(spoken in southern India, Sri Lanka, and Singapore), Telugu and Malayalam (spoken mainly in southern India), and the
Niger-Congo languages Swahili and Yoruba, spoken in Africa.

• Лицензия : Нет известной лицензии.
• Версия : 1.0.0
• Расколы :

• Функции :

{
    "tokens": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "pos_tags": {
        "feature": {
            "num_classes": 17,
            "names": [
                "ADJ",
                "ADP",
                "ADV",
                "AUX",
                "CCONJ",
                "DET",
                "INTJ",
                "NOUN",
                "NUM",
                "PART",
                "PRON",
                "PROPN",
                "PUNCT",
                "SCONJ",
                "SYM",
                "VERB",
                "X"
            ],
            "names_file": null,
            "id": null,
            "_type": "ClassLabel"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

udpos.Тайский

Используйте следующую команду, чтобы загрузить этот набор данных в TFDS:

ds = tfds.load('huggingface:xtreme/udpos.Thai')

• Описание :

Universal Dependencies (UD) is a framework for consistent annotation of grammar (parts of speech, morphological
features, and syntactic dependencies) across different human languages. UD is an open community effort with over 200
contributors producing more than 100 treebanks in over 70 languages. If you’re new to UD, you should start by reading
the first part of the Short Introduction and then browsing the annotation guidelines.

The Cross-lingual TRansfer Evaluation of Multilingual Encoders (XTREME) benchmark is a benchmark for the evaluation of
the cross-lingual generalization ability of pre-trained multilingual models. It covers 40 typologically diverse languages
(spanning 12 language families) and includes nine tasks that collectively require reasoning about different levels of
syntax and semantics. The languages in XTREME are selected to maximize language diversity, coverage in existing tasks,
and availability of training data. Among these are many under-studied languages, such as the Dravidian languages Tamil
(spoken in southern India, Sri Lanka, and Singapore), Telugu and Malayalam (spoken mainly in southern India), and the
Niger-Congo languages Swahili and Yoruba, spoken in Africa.

• Лицензия : Нет известной лицензии.
• Версия : 1.0.0
• Расколы :

• Функции :

{
    "tokens": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "pos_tags": {
        "feature": {
            "num_classes": 17,
            "names": [
                "ADJ",
                "ADP",
                "ADV",
                "AUX",
                "CCONJ",
                "DET",
                "INTJ",
                "NOUN",
                "NUM",
                "PART",
                "PRON",
                "PROPN",
                "PUNCT",
                "SCONJ",
                "SYM",
                "VERB",
                "X"
            ],
            "names_file": null,
            "id": null,
            "_type": "ClassLabel"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

udpos.турецкий

Используйте следующую команду, чтобы загрузить этот набор данных в TFDS:

ds = tfds.load('huggingface:xtreme/udpos.Turkish')

• Описание :

Universal Dependencies (UD) is a framework for consistent annotation of grammar (parts of speech, morphological
features, and syntactic dependencies) across different human languages. UD is an open community effort with over 200
contributors producing more than 100 treebanks in over 70 languages. If you’re new to UD, you should start by reading
the first part of the Short Introduction and then browsing the annotation guidelines.

The Cross-lingual TRansfer Evaluation of Multilingual Encoders (XTREME) benchmark is a benchmark for the evaluation of
the cross-lingual generalization ability of pre-trained multilingual models. It covers 40 typologically diverse languages
(spanning 12 language families) and includes nine tasks that collectively require reasoning about different levels of
syntax and semantics. The languages in XTREME are selected to maximize language diversity, coverage in existing tasks,
and availability of training data. Among these are many under-studied languages, such as the Dravidian languages Tamil
(spoken in southern India, Sri Lanka, and Singapore), Telugu and Malayalam (spoken mainly in southern India), and the
Niger-Congo languages Swahili and Yoruba, spoken in Africa.

• Лицензия : Нет известной лицензии.
• Версия : 1.0.0
• Расколы :

• Функции :

{
    "tokens": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "pos_tags": {
        "feature": {
            "num_classes": 17,
            "names": [
                "ADJ",
                "ADP",
                "ADV",
                "AUX",
                "CCONJ",
                "DET",
                "INTJ",
                "NOUN",
                "NUM",
                "PART",
                "PRON",
                "PROPN",
                "PUNCT",
                "SCONJ",
                "SYM",
                "VERB",
                "X"
            ],
            "names_file": null,
            "id": null,
            "_type": "ClassLabel"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

udpos.урду

Используйте следующую команду, чтобы загрузить этот набор данных в TFDS:

ds = tfds.load('huggingface:xtreme/udpos.Urdu')

• Описание :

Universal Dependencies (UD) is a framework for consistent annotation of grammar (parts of speech, morphological
features, and syntactic dependencies) across different human languages. UD is an open community effort with over 200
contributors producing more than 100 treebanks in over 70 languages. If you’re new to UD, you should start by reading
the first part of the Short Introduction and then browsing the annotation guidelines.

The Cross-lingual TRansfer Evaluation of Multilingual Encoders (XTREME) benchmark is a benchmark for the evaluation of
the cross-lingual generalization ability of pre-trained multilingual models. It covers 40 typologically diverse languages
(spanning 12 language families) and includes nine tasks that collectively require reasoning about different levels of
syntax and semantics. The languages in XTREME are selected to maximize language diversity, coverage in existing tasks,
and availability of training data. Among these are many under-studied languages, such as the Dravidian languages Tamil
(spoken in southern India, Sri Lanka, and Singapore), Telugu and Malayalam (spoken mainly in southern India), and the
Niger-Congo languages Swahili and Yoruba, spoken in Africa.

• Лицензия : Нет известной лицензии.
• Версия : 1.0.0
• Расколы :

• Функции :

{
    "tokens": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "pos_tags": {
        "feature": {
            "num_classes": 17,
            "names": [
                "ADJ",
                "ADP",
                "ADV",
                "AUX",
                "CCONJ",
                "DET",
                "INTJ",
                "NOUN",
                "NUM",
                "PART",
                "PRON",
                "PROPN",
                "PUNCT",
                "SCONJ",
                "SYM",
                "VERB",
                "X"
            ],
            "names_file": null,
            "id": null,
            "_type": "ClassLabel"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

udpos.Вьетнамский

Используйте следующую команду, чтобы загрузить этот набор данных в TFDS:

ds = tfds.load('huggingface:xtreme/udpos.Vietnamese')

• Описание :

Universal Dependencies (UD) is a framework for consistent annotation of grammar (parts of speech, morphological
features, and syntactic dependencies) across different human languages. UD is an open community effort with over 200
contributors producing more than 100 treebanks in over 70 languages. If you’re new to UD, you should start by reading
the first part of the Short Introduction and then browsing the annotation guidelines.

The Cross-lingual TRansfer Evaluation of Multilingual Encoders (XTREME) benchmark is a benchmark for the evaluation of
the cross-lingual generalization ability of pre-trained multilingual models. It covers 40 typologically diverse languages
(spanning 12 language families) and includes nine tasks that collectively require reasoning about different levels of
syntax and semantics. The languages in XTREME are selected to maximize language diversity, coverage in existing tasks,
and availability of training data. Among these are many under-studied languages, such as the Dravidian languages Tamil
(spoken in southern India, Sri Lanka, and Singapore), Telugu and Malayalam (spoken mainly in southern India), and the
Niger-Congo languages Swahili and Yoruba, spoken in Africa.

• Лицензия : Нет известной лицензии.
• Версия : 1.0.0
• Расколы :

• Функции :

{
    "tokens": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "pos_tags": {
        "feature": {
            "num_classes": 17,
            "names": [
                "ADJ",
                "ADP",
                "ADV",
                "AUX",
                "CCONJ",
                "DET",
                "INTJ",
                "NOUN",
                "NUM",
                "PART",
                "PRON",
                "PROPN",
                "PUNCT",
                "SCONJ",
                "SYM",
                "VERB",
                "X"
            ],
            "names_file": null,
            "id": null,
            "_type": "ClassLabel"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

удпос.Йоруба

Используйте следующую команду, чтобы загрузить этот набор данных в TFDS:

ds = tfds.load('huggingface:xtreme/udpos.Yoruba')

• Описание :

Universal Dependencies (UD) is a framework for consistent annotation of grammar (parts of speech, morphological
features, and syntactic dependencies) across different human languages. UD is an open community effort with over 200
contributors producing more than 100 treebanks in over 70 languages. If you’re new to UD, you should start by reading
the first part of the Short Introduction and then browsing the annotation guidelines.

The Cross-lingual TRansfer Evaluation of Multilingual Encoders (XTREME) benchmark is a benchmark for the evaluation of
the cross-lingual generalization ability of pre-trained multilingual models. It covers 40 typologically diverse languages
(spanning 12 language families) and includes nine tasks that collectively require reasoning about different levels of
syntax and semantics. The languages in XTREME are selected to maximize language diversity, coverage in existing tasks,
and availability of training data. Among these are many under-studied languages, such as the Dravidian languages Tamil
(spoken in southern India, Sri Lanka, and Singapore), Telugu and Malayalam (spoken mainly in southern India), and the
Niger-Congo languages Swahili and Yoruba, spoken in Africa.

• Лицензия : Нет известной лицензии.
• Версия : 1.0.0
• Расколы :

• Функции :

{
    "tokens": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "pos_tags": {
        "feature": {
            "num_classes": 17,
            "names": [
                "ADJ",
                "ADP",
                "ADV",
                "AUX",
                "CCONJ",
                "DET",
                "INTJ",
                "NOUN",
                "NUM",
                "PART",
                "PRON",
                "PROPN",
                "PUNCT",
                "SCONJ",
                "SYM",
                "VERB",
                "X"
            ],
            "names_file": null,
            "id": null,
            "_type": "ClassLabel"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}