большая скамейка

Ссылки:

абстрактное_нарративное_понимание

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

ds = tfds.load('huggingface:bigbench/abstract_narrative_understanding')
  • Описание :
The Beyond the Imitation Game Benchmark (BIG-bench) is a collaborative benchmark intended to
probe large language models, and extrapolate their future capabilities.
  • Лицензия : Лицензия Apache 2.0.
  • Версия : 0.0.0
  • Расколы :
Расколоть Примеры
'default' 3000
'train' 2400
'validation' 600
  • Функции :
{
    "idx": {
        "dtype": "int32",
        "id": null,
        "_type": "Value"
    },
    "inputs": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_scores": {
        "feature": {
            "dtype": "int32",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

анахронизмы

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

ds = tfds.load('huggingface:bigbench/anachronisms')
  • Описание :
The Beyond the Imitation Game Benchmark (BIG-bench) is a collaborative benchmark intended to
probe large language models, and extrapolate their future capabilities.
  • Лицензия : Лицензия Apache 2.0.
  • Версия : 0.0.0
  • Расколы :
Расколоть Примеры
'default' 230
'train' 184
'validation' 46
  • Функции :
{
    "idx": {
        "dtype": "int32",
        "id": null,
        "_type": "Value"
    },
    "inputs": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_scores": {
        "feature": {
            "dtype": "int32",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

аналогическое_сходство

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

ds = tfds.load('huggingface:bigbench/analogical_similarity')
  • Описание :
The Beyond the Imitation Game Benchmark (BIG-bench) is a collaborative benchmark intended to
probe large language models, and extrapolate their future capabilities.
  • Лицензия : Лицензия Apache 2.0.
  • Версия : 0.0.0
  • Расколы :
Расколоть Примеры
'default' 323
'train' 259
'validation' 64
  • Функции :
{
    "idx": {
        "dtype": "int32",
        "id": null,
        "_type": "Value"
    },
    "inputs": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_scores": {
        "feature": {
            "dtype": "int32",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

analytics_entailment

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

ds = tfds.load('huggingface:bigbench/analytic_entailment')
  • Описание :
The Beyond the Imitation Game Benchmark (BIG-bench) is a collaborative benchmark intended to
probe large language models, and extrapolate their future capabilities.
  • Лицензия : Лицензия Apache 2.0.
  • Версия : 0.0.0
  • Расколы :
Расколоть Примеры
'default' 70
'train' 54
'validation' 16
  • Функции :
{
    "idx": {
        "dtype": "int32",
        "id": null,
        "_type": "Value"
    },
    "inputs": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_scores": {
        "feature": {
            "dtype": "int32",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

арифметика

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

ds = tfds.load('huggingface:bigbench/arithmetic')
  • Описание :
The Beyond the Imitation Game Benchmark (BIG-bench) is a collaborative benchmark intended to
probe large language models, and extrapolate their future capabilities.
  • Лицензия : Лицензия Apache 2.0.
  • Версия : 0.0.0
  • Расколы :
Расколоть Примеры
'default' 15023
'train' 12019
'validation' 3004
  • Функции :
{
    "idx": {
        "dtype": "int32",
        "id": null,
        "_type": "Value"
    },
    "inputs": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_scores": {
        "feature": {
            "dtype": "int32",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

ascii_word_recognition

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

ds = tfds.load('huggingface:bigbench/ascii_word_recognition')
  • Описание :
The Beyond the Imitation Game Benchmark (BIG-bench) is a collaborative benchmark intended to
probe large language models, and extrapolate their future capabilities.
  • Лицензия : Лицензия Apache 2.0.
  • Версия : 0.0.0
  • Расколы :
Расколоть Примеры
'default' 5000
'train' 4000
'validation' 1000
  • Функции :
{
    "idx": {
        "dtype": "int32",
        "id": null,
        "_type": "Value"
    },
    "inputs": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_scores": {
        "feature": {
            "dtype": "int32",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

авторство_проверка

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

ds = tfds.load('huggingface:bigbench/authorship_verification')
  • Описание :
The Beyond the Imitation Game Benchmark (BIG-bench) is a collaborative benchmark intended to
probe large language models, and extrapolate their future capabilities.
  • Лицензия : Лицензия Apache 2.0.
  • Версия : 0.0.0
  • Расколы :
Расколоть Примеры
'default' 880
'train' 704
'validation' 176
  • Функции :
{
    "idx": {
        "dtype": "int32",
        "id": null,
        "_type": "Value"
    },
    "inputs": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_scores": {
        "feature": {
            "dtype": "int32",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

auto_categorization

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

ds = tfds.load('huggingface:bigbench/auto_categorization')
  • Описание :
The Beyond the Imitation Game Benchmark (BIG-bench) is a collaborative benchmark intended to
probe large language models, and extrapolate their future capabilities.
  • Лицензия : Лицензия Apache 2.0.
  • Версия : 0.0.0
  • Расколы :
Расколоть Примеры
'default' 328
'train' 263
'validation' 65
  • Функции :
{
    "idx": {
        "dtype": "int32",
        "id": null,
        "_type": "Value"
    },
    "inputs": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_scores": {
        "feature": {
            "dtype": "int32",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

auto_debugging

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

ds = tfds.load('huggingface:bigbench/auto_debugging')
  • Описание :
The Beyond the Imitation Game Benchmark (BIG-bench) is a collaborative benchmark intended to
probe large language models, and extrapolate their future capabilities.
  • Лицензия : Лицензия Apache 2.0.
  • Версия : 0.0.0
  • Расколы :
Расколоть Примеры
'default' 34
'train' 18
'validation' 16
  • Функции :
{
    "idx": {
        "dtype": "int32",
        "id": null,
        "_type": "Value"
    },
    "inputs": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_scores": {
        "feature": {
            "dtype": "int32",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

bbq_lite_json

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

ds = tfds.load('huggingface:bigbench/bbq_lite_json')
  • Описание :
The Beyond the Imitation Game Benchmark (BIG-bench) is a collaborative benchmark intended to
probe large language models, and extrapolate their future capabilities.
  • Лицензия : Лицензия Apache 2.0.
  • Версия : 0.0.0
  • Расколы :
Расколоть Примеры
'default' 16076
'train' 12866
'validation' 3210
  • Функции :
{
    "idx": {
        "dtype": "int32",
        "id": null,
        "_type": "Value"
    },
    "inputs": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_scores": {
        "feature": {
            "dtype": "int32",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

bridging_anaphora_solve_barqa

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

ds = tfds.load('huggingface:bigbench/bridging_anaphora_resolution_barqa')
  • Описание :
The Beyond the Imitation Game Benchmark (BIG-bench) is a collaborative benchmark intended to
probe large language models, and extrapolate their future capabilities.
  • Лицензия : Лицензия Apache 2.0.
  • Версия : 0.0.0
  • Расколы :
Расколоть Примеры
'default' 648
'train' 519
'validation' 129
  • Функции :
{
    "idx": {
        "dtype": "int32",
        "id": null,
        "_type": "Value"
    },
    "inputs": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_scores": {
        "feature": {
            "dtype": "int32",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

причинное_суждение

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

ds = tfds.load('huggingface:bigbench/causal_judgment')
  • Описание :
The Beyond the Imitation Game Benchmark (BIG-bench) is a collaborative benchmark intended to
probe large language models, and extrapolate their future capabilities.
  • Лицензия : Лицензия Apache 2.0.
  • Версия : 0.0.0
  • Расколы :
Расколоть Примеры
'default' 190
'train' 152
'validation' 38
  • Функции :
{
    "idx": {
        "dtype": "int32",
        "id": null,
        "_type": "Value"
    },
    "inputs": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_scores": {
        "feature": {
            "dtype": "int32",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

причина_и_следствие

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

ds = tfds.load('huggingface:bigbench/cause_and_effect')
  • Описание :
The Beyond the Imitation Game Benchmark (BIG-bench) is a collaborative benchmark intended to
probe large language models, and extrapolate their future capabilities.
  • Лицензия : Лицензия Apache 2.0.
  • Версия : 0.0.0
  • Расколы :
Расколоть Примеры
'default' 153
'train' 123
'validation' 30
  • Функции :
{
    "idx": {
        "dtype": "int32",
        "id": null,
        "_type": "Value"
    },
    "inputs": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_scores": {
        "feature": {
            "dtype": "int32",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

checkmate_in_one

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

ds = tfds.load('huggingface:bigbench/checkmate_in_one')
  • Описание :
The Beyond the Imitation Game Benchmark (BIG-bench) is a collaborative benchmark intended to
probe large language models, and extrapolate their future capabilities.
  • Лицензия : Лицензия Apache 2.0.
  • Версия : 0.0.0
  • Расколы :
Расколоть Примеры
'default' 3498
'train' 2799
'validation' 699
  • Функции :
{
    "idx": {
        "dtype": "int32",
        "id": null,
        "_type": "Value"
    },
    "inputs": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_scores": {
        "feature": {
            "dtype": "int32",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

Chess_state_tracking

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

ds = tfds.load('huggingface:bigbench/chess_state_tracking')
  • Описание :
The Beyond the Imitation Game Benchmark (BIG-bench) is a collaborative benchmark intended to
probe large language models, and extrapolate their future capabilities.
  • Лицензия : Лицензия Apache 2.0.
  • Версия : 0.0.0
  • Расколы :
Расколоть Примеры
'default' 6000
'train' 4800
'validation' 1200
  • Функции :
{
    "idx": {
        "dtype": "int32",
        "id": null,
        "_type": "Value"
    },
    "inputs": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_scores": {
        "feature": {
            "dtype": "int32",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

chinese_remainder_theorem

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

ds = tfds.load('huggingface:bigbench/chinese_remainder_theorem')
  • Описание :
The Beyond the Imitation Game Benchmark (BIG-bench) is a collaborative benchmark intended to
probe large language models, and extrapolate their future capabilities.
  • Лицензия : Лицензия Apache 2.0.
  • Версия : 0.0.0
  • Расколы :
Расколоть Примеры
'default' 500
'train' 400
'validation' 100
  • Функции :
{
    "idx": {
        "dtype": "int32",
        "id": null,
        "_type": "Value"
    },
    "inputs": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_scores": {
        "feature": {
            "dtype": "int32",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

cifar10_classification

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

ds = tfds.load('huggingface:bigbench/cifar10_classification')
  • Описание :
The Beyond the Imitation Game Benchmark (BIG-bench) is a collaborative benchmark intended to
probe large language models, and extrapolate their future capabilities.
  • Лицензия : Лицензия Apache 2.0.
  • Версия : 0.0.0
  • Расколы :
Расколоть Примеры
'default' 20000
'train' 16000
'validation' 4000
  • Функции :
{
    "idx": {
        "dtype": "int32",
        "id": null,
        "_type": "Value"
    },
    "inputs": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_scores": {
        "feature": {
            "dtype": "int32",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

code_line_description

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

ds = tfds.load('huggingface:bigbench/code_line_description')
  • Описание :
The Beyond the Imitation Game Benchmark (BIG-bench) is a collaborative benchmark intended to
probe large language models, and extrapolate their future capabilities.
  • Лицензия : Лицензия Apache 2.0.
  • Версия : 0.0.0
  • Расколы :
Расколоть Примеры
'default' 60
'train' 44
'validation' 16
  • Функции :
{
    "idx": {
        "dtype": "int32",
        "id": null,
        "_type": "Value"
    },
    "inputs": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_scores": {
        "feature": {
            "dtype": "int32",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

кодовые имена

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

ds = tfds.load('huggingface:bigbench/codenames')
  • Описание :
The Beyond the Imitation Game Benchmark (BIG-bench) is a collaborative benchmark intended to
probe large language models, and extrapolate their future capabilities.
  • Лицензия : Лицензия Apache 2.0.
  • Версия : 0.0.0
  • Расколы :
Расколоть Примеры
'default' 85
'train' 68
'validation' 17
  • Функции :
{
    "idx": {
        "dtype": "int32",
        "id": null,
        "_type": "Value"
    },
    "inputs": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_scores": {
        "feature": {
            "dtype": "int32",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

цвет

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

ds = tfds.load('huggingface:bigbench/color')
  • Описание :
The Beyond the Imitation Game Benchmark (BIG-bench) is a collaborative benchmark intended to
probe large language models, and extrapolate their future capabilities.
  • Лицензия : Лицензия Apache 2.0.
  • Версия : 0.0.0
  • Расколы :
Расколоть Примеры
'default' 4000
'train' 3200
'validation' 800
  • Функции :
{
    "idx": {
        "dtype": "int32",
        "id": null,
        "_type": "Value"
    },
    "inputs": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_scores": {
        "feature": {
            "dtype": "int32",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

common_morpheme

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

ds = tfds.load('huggingface:bigbench/common_morpheme')
  • Описание :
The Beyond the Imitation Game Benchmark (BIG-bench) is a collaborative benchmark intended to
probe large language models, and extrapolate their future capabilities.
  • Лицензия : Лицензия Apache 2.0.
  • Версия : 0.0.0
  • Расколы :
Расколоть Примеры
'default' 50
'train' 34
'validation' 16
  • Функции :
{
    "idx": {
        "dtype": "int32",
        "id": null,
        "_type": "Value"
    },
    "inputs": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_scores": {
        "feature": {
            "dtype": "int32",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

концептуальные_комбинации

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

ds = tfds.load('huggingface:bigbench/conceptual_combinations')
  • Описание :
The Beyond the Imitation Game Benchmark (BIG-bench) is a collaborative benchmark intended to
probe large language models, and extrapolate their future capabilities.
  • Лицензия : Лицензия Apache 2.0.
  • Версия : 0.0.0
  • Расколы :
Расколоть Примеры
'default' 103
'train' 84
'validation' 19
  • Функции :
{
    "idx": {
        "dtype": "int32",
        "id": null,
        "_type": "Value"
    },
    "inputs": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_scores": {
        "feature": {
            "dtype": "int32",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

conlang_translation

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

ds = tfds.load('huggingface:bigbench/conlang_translation')
  • Описание :
The Beyond the Imitation Game Benchmark (BIG-bench) is a collaborative benchmark intended to
probe large language models, and extrapolate their future capabilities.
  • Лицензия : Лицензия Apache 2.0.
  • Версия : 0.0.0
  • Расколы :
Расколоть Примеры
'default' 164
'train' 132
'validation' 32
  • Функции :
{
    "idx": {
        "dtype": "int32",
        "id": null,
        "_type": "Value"
    },
    "inputs": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_scores": {
        "feature": {
            "dtype": "int32",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

contextual_parametric_knowledge_conflicts

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

ds = tfds.load('huggingface:bigbench/contextual_parametric_knowledge_conflicts')
  • Описание :
The Beyond the Imitation Game Benchmark (BIG-bench) is a collaborative benchmark intended to
probe large language models, and extrapolate their future capabilities.
  • Лицензия : Лицензия Apache 2.0.
  • Версия : 0.0.0
  • Расколы :
Расколоть Примеры
'default' 17528
'train' 14023
'validation' 3505
  • Функции :
{
    "idx": {
        "dtype": "int32",
        "id": null,
        "_type": "Value"
    },
    "inputs": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_scores": {
        "feature": {
            "dtype": "int32",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

Crash_blossom

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

ds = tfds.load('huggingface:bigbench/crash_blossom')
  • Описание :
The Beyond the Imitation Game Benchmark (BIG-bench) is a collaborative benchmark intended to
probe large language models, and extrapolate their future capabilities.
  • Лицензия : Лицензия Apache 2.0.
  • Версия : 0.0.0
  • Расколы :
Расколоть Примеры
'default' 38
'train' 22
'validation' 16
  • Функции :
{
    "idx": {
        "dtype": "int32",
        "id": null,
        "_type": "Value"
    },
    "inputs": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_scores": {
        "feature": {
            "dtype": "int32",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

грубый_ай

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

ds = tfds.load('huggingface:bigbench/crass_ai')
  • Описание :
The Beyond the Imitation Game Benchmark (BIG-bench) is a collaborative benchmark intended to
probe large language models, and extrapolate their future capabilities.
  • Лицензия : Лицензия Apache 2.0.
  • Версия : 0.0.0
  • Расколы :
Расколоть Примеры
'default' 44
'train' 28
'validation' 16
  • Функции :
{
    "idx": {
        "dtype": "int32",
        "id": null,
        "_type": "Value"
    },
    "inputs": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_scores": {
        "feature": {
            "dtype": "int32",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

криобиология_испанский

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

ds = tfds.load('huggingface:bigbench/cryobiology_spanish')
  • Описание :
The Beyond the Imitation Game Benchmark (BIG-bench) is a collaborative benchmark intended to
probe large language models, and extrapolate their future capabilities.
  • Лицензия : Лицензия Apache 2.0.
  • Версия : 0.0.0
  • Расколы :
Расколоть Примеры
'default' 146
'train' 117
'validation' 29
  • Функции :
{
    "idx": {
        "dtype": "int32",
        "id": null,
        "_type": "Value"
    },
    "inputs": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_scores": {
        "feature": {
            "dtype": "int32",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

криптонит

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

ds = tfds.load('huggingface:bigbench/cryptonite')
  • Описание :
The Beyond the Imitation Game Benchmark (BIG-bench) is a collaborative benchmark intended to
probe large language models, and extrapolate their future capabilities.
  • Лицензия : Лицензия Apache 2.0.
  • Версия : 0.0.0
  • Расколы :
Расколоть Примеры
'default' 26157
'train' 20926
'validation' 5231
  • Функции :
{
    "idx": {
        "dtype": "int32",
        "id": null,
        "_type": "Value"
    },
    "inputs": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_scores": {
        "feature": {
            "dtype": "int32",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

cs_algorithms

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

ds = tfds.load('huggingface:bigbench/cs_algorithms')
  • Описание :
The Beyond the Imitation Game Benchmark (BIG-bench) is a collaborative benchmark intended to
probe large language models, and extrapolate their future capabilities.
  • Лицензия : Лицензия Apache 2.0.
  • Версия : 0.0.0
  • Расколы :
Расколоть Примеры
'default' 1320
'train' 1056
'validation' 264
  • Функции :
{
    "idx": {
        "dtype": "int32",
        "id": null,
        "_type": "Value"
    },
    "inputs": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_scores": {
        "feature": {
            "dtype": "int32",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

dark_humor_detection

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

ds = tfds.load('huggingface:bigbench/dark_humor_detection')
  • Описание :
The Beyond the Imitation Game Benchmark (BIG-bench) is a collaborative benchmark intended to
probe large language models, and extrapolate their future capabilities.
  • Лицензия : Лицензия Apache 2.0.
  • Версия : 0.0.0
  • Расколы :
Расколоть Примеры
'default' 80
'train' 64
'validation' 16
  • Функции :
{
    "idx": {
        "dtype": "int32",
        "id": null,
        "_type": "Value"
    },
    "inputs": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_scores": {
        "feature": {
            "dtype": "int32",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

дата_понимание

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

ds = tfds.load('huggingface:bigbench/date_understanding')
  • Описание :
The Beyond the Imitation Game Benchmark (BIG-bench) is a collaborative benchmark intended to
probe large language models, and extrapolate their future capabilities.
  • Лицензия : Лицензия Apache 2.0.
  • Версия : 0.0.0
  • Расколы :
Расколоть Примеры
'default' 369
'train' 296
'validation' 73
  • Функции :
{
    "idx": {
        "dtype": "int32",
        "id": null,
        "_type": "Value"
    },
    "inputs": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_scores": {
        "feature": {
            "dtype": "int32",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

значения_qa

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

ds = tfds.load('huggingface:bigbench/disambiguation_qa')
  • Описание :
The Beyond the Imitation Game Benchmark (BIG-bench) is a collaborative benchmark intended to
probe large language models, and extrapolate their future capabilities.
  • Лицензия : Лицензия Apache 2.0.
  • Версия : 0.0.0
  • Расколы :
Расколоть Примеры
'default' 258
'train' 207
'validation' 51
  • Функции :
{
    "idx": {
        "dtype": "int32",
        "id": null,
        "_type": "Value"
    },
    "inputs": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_scores": {
        "feature": {
            "dtype": "int32",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

discourse_marker_prediction

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

ds = tfds.load('huggingface:bigbench/discourse_marker_prediction')
  • Описание :
The Beyond the Imitation Game Benchmark (BIG-bench) is a collaborative benchmark intended to
probe large language models, and extrapolate their future capabilities.
  • Лицензия : Лицензия Apache 2.0.
  • Версия : 0.0.0
  • Расколы :
Расколоть Примеры
'default' 857
'train' 686
'validation' 171
  • Функции :
{
    "idx": {
        "dtype": "int32",
        "id": null,
        "_type": "Value"
    },
    "inputs": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_scores": {
        "feature": {
            "dtype": "int32",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

disfl_qa

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

ds = tfds.load('huggingface:bigbench/disfl_qa')
  • Описание :
The Beyond the Imitation Game Benchmark (BIG-bench) is a collaborative benchmark intended to
probe large language models, and extrapolate their future capabilities.
  • Лицензия : Лицензия Apache 2.0.
  • Версия : 0.0.0
  • Расколы :
Расколоть Примеры
'default' 8000
'train' 6400
'validation' 1600
  • Функции :
{
    "idx": {
        "dtype": "int32",
        "id": null,
        "_type": "Value"
    },
    "inputs": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_scores": {
        "feature": {
            "dtype": "int32",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

dyck_languages

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

ds = tfds.load('huggingface:bigbench/dyck_languages')
  • Описание :
The Beyond the Imitation Game Benchmark (BIG-bench) is a collaborative benchmark intended to
probe large language models, and extrapolate their future capabilities.
  • Лицензия : Лицензия Apache 2.0.
  • Версия : 0.0.0
  • Расколы :
Расколоть Примеры
'default' 1000
'train' 800
'validation' 200
  • Функции :
{
    "idx": {
        "dtype": "int32",
        "id": null,
        "_type": "Value"
    },
    "inputs": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_scores": {
        "feature": {
            "dtype": "int32",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

элементарный_math_qa

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

ds = tfds.load('huggingface:bigbench/elementary_math_qa')
  • Описание :
The Beyond the Imitation Game Benchmark (BIG-bench) is a collaborative benchmark intended to
probe large language models, and extrapolate their future capabilities.
  • Лицензия : Лицензия Apache 2.0.
  • Версия : 0.0.0
  • Расколы :
Расколоть Примеры
'default' 38160
'train' 30531
'validation' 7629
  • Функции :
{
    "idx": {
        "dtype": "int32",
        "id": null,
        "_type": "Value"
    },
    "inputs": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_scores": {
        "feature": {
            "dtype": "int32",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

emoji_movie

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

ds = tfds.load('huggingface:bigbench/emoji_movie')
  • Описание :
The Beyond the Imitation Game Benchmark (BIG-bench) is a collaborative benchmark intended to
probe large language models, and extrapolate their future capabilities.
  • Лицензия : Лицензия Apache 2.0.
  • Версия : 0.0.0
  • Расколы :
Расколоть Примеры
'default' 100
'train' 80
'validation' 20
  • Функции :
{
    "idx": {
        "dtype": "int32",
        "id": null,
        "_type": "Value"
    },
    "inputs": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_scores": {
        "feature": {
            "dtype": "int32",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

emojis_emotion_prediction

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

ds = tfds.load('huggingface:bigbench/emojis_emotion_prediction')
  • Описание :
The Beyond the Imitation Game Benchmark (BIG-bench) is a collaborative benchmark intended to
probe large language models, and extrapolate their future capabilities.
  • Лицензия : Лицензия Apache 2.0.
  • Версия : 0.0.0
  • Расколы :
Расколоть Примеры
'default' 131
'train' 105
'validation' 26
  • Функции :
{
    "idx": {
        "dtype": "int32",
        "id": null,
        "_type": "Value"
    },
    "inputs": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_scores": {
        "feature": {
            "dtype": "int32",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

эмпирические_суждения

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

ds = tfds.load('huggingface:bigbench/empirical_judgments')
  • Описание :
The Beyond the Imitation Game Benchmark (BIG-bench) is a collaborative benchmark intended to
probe large language models, and extrapolate their future capabilities.
  • Лицензия : Лицензия Apache 2.0.
  • Версия : 0.0.0
  • Расколы :
Расколоть Примеры
'default' 99
'train' 80
'validation' 19
  • Функции :
{
    "idx": {
        "dtype": "int32",
        "id": null,
        "_type": "Value"
    },
    "inputs": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_scores": {
        "feature": {
            "dtype": "int32",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

английские_пословицы

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

ds = tfds.load('huggingface:bigbench/english_proverbs')
  • Описание :
The Beyond the Imitation Game Benchmark (BIG-bench) is a collaborative benchmark intended to
probe large language models, and extrapolate their future capabilities.
  • Лицензия : Лицензия Apache 2.0.
  • Версия : 0.0.0
  • Расколы :
Расколоть Примеры
'default' 34
'train' 18
'validation' 16
  • Функции :
{
    "idx": {
        "dtype": "int32",
        "id": null,
        "_type": "Value"
    },
    "inputs": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_scores": {
        "feature": {
            "dtype": "int32",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

английские_русские_пословицы

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

ds = tfds.load('huggingface:bigbench/english_russian_proverbs')
  • Описание :
The Beyond the Imitation Game Benchmark (BIG-bench) is a collaborative benchmark intended to
probe large language models, and extrapolate their future capabilities.
  • Лицензия : Лицензия Apache 2.0.
  • Версия : 0.0.0
  • Расколы :
Расколоть Примеры
'default' 80
'train' 64
'validation' 16
  • Функции :
{
    "idx": {
        "dtype": "int32",
        "id": null,
        "_type": "Value"
    },
    "inputs": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_scores": {
        "feature": {
            "dtype": "int32",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

entailed_polarity

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

ds = tfds.load('huggingface:bigbench/entailed_polarity')
  • Описание :
The Beyond the Imitation Game Benchmark (BIG-bench) is a collaborative benchmark intended to
probe large language models, and extrapolate their future capabilities.
  • Лицензия : Лицензия Apache 2.0.
  • Версия : 0.0.0
  • Расколы :
Расколоть Примеры
'default' 148
'train' 119
'validation' 29
  • Функции :
{
    "idx": {
        "dtype": "int32",
        "id": null,
        "_type": "Value"
    },
    "inputs": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_scores": {
        "feature": {
            "dtype": "int32",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

entailed_polarity_hindi

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

ds = tfds.load('huggingface:bigbench/entailed_polarity_hindi')
  • Описание :
The Beyond the Imitation Game Benchmark (BIG-bench) is a collaborative benchmark intended to
probe large language models, and extrapolate their future capabilities.
  • Лицензия : Лицензия Apache 2.0.
  • Версия : 0.0.0
  • Расколы :
Расколоть Примеры
'default' 138
'train' 111
'validation' 27
  • Функции :
{
    "idx": {
        "dtype": "int32",
        "id": null,
        "_type": "Value"
    },
    "inputs": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_scores": {
        "feature": {
            "dtype": "int32",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

epistemic_reasoning

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

ds = tfds.load('huggingface:bigbench/epistemic_reasoning')
  • Описание :
The Beyond the Imitation Game Benchmark (BIG-bench) is a collaborative benchmark intended to
probe large language models, and extrapolate their future capabilities.
  • Лицензия : Лицензия Apache 2.0.
  • Версия : 0.0.0
  • Расколы :
Расколоть Примеры
'default' 2000 г.
'train' 1600
'validation' 400
  • Функции :
{
    "idx": {
        "dtype": "int32",
        "id": null,
        "_type": "Value"
    },
    "inputs": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_scores": {
        "feature": {
            "dtype": "int32",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

оценивающая_информация_существенность

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

ds = tfds.load('huggingface:bigbench/evaluating_information_essentiality')
  • Описание :
The Beyond the Imitation Game Benchmark (BIG-bench) is a collaborative benchmark intended to
probe large language models, and extrapolate their future capabilities.
  • Лицензия : Лицензия Apache 2.0.
  • Версия : 0.0.0
  • Расколы :
Расколоть Примеры
'default' 68
'train' 52
'validation' 16
  • Функции :
{
    "idx": {
        "dtype": "int32",
        "id": null,
        "_type": "Value"
    },
    "inputs": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_scores": {
        "feature": {
            "dtype": "int32",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

факт_проверка

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

ds = tfds.load('huggingface:bigbench/fact_checker')
  • Описание :
The Beyond the Imitation Game Benchmark (BIG-bench) is a collaborative benchmark intended to
probe large language models, and extrapolate their future capabilities.
  • Лицензия : Лицензия Apache 2.0.
  • Версия : 0.0.0
  • Расколы :
Расколоть Примеры
'default' 7154
'train' 5724
'validation' 14:30
  • Функции :
{
    "idx": {
        "dtype": "int32",
        "id": null,
        "_type": "Value"
    },
    "inputs": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_scores": {
        "feature": {
            "dtype": "int32",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

Fantasy_reasoning

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

ds = tfds.load('huggingface:bigbench/fantasy_reasoning')
  • Описание :
The Beyond the Imitation Game Benchmark (BIG-bench) is a collaborative benchmark intended to
probe large language models, and extrapolate their future capabilities.
  • Лицензия : Лицензия Apache 2.0.
  • Версия : 0.0.0
  • Расколы :
Расколоть Примеры
'default' 201
'train' 161
'validation' 40
  • Функции :
{
    "idx": {
        "dtype": "int32",
        "id": null,
        "_type": "Value"
    },
    "inputs": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_scores": {
        "feature": {
            "dtype": "int32",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

несколько_shot_nlg

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

ds = tfds.load('huggingface:bigbench/few_shot_nlg')
  • Описание :
The Beyond the Imitation Game Benchmark (BIG-bench) is a collaborative benchmark intended to
probe large language models, and extrapolate their future capabilities.
  • Лицензия : Лицензия Apache 2.0.
  • Версия : 0.0.0
  • Расколы :
Расколоть Примеры
'default' 153
'train' 123
'validation' 30
  • Функции :
{
    "idx": {
        "dtype": "int32",
        "id": null,
        "_type": "Value"
    },
    "inputs": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_scores": {
        "feature": {
            "dtype": "int32",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

фигура_of_speech_detection

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

ds = tfds.load('huggingface:bigbench/figure_of_speech_detection')
  • Описание :
The Beyond the Imitation Game Benchmark (BIG-bench) is a collaborative benchmark intended to
probe large language models, and extrapolate their future capabilities.
  • Лицензия : Лицензия Apache 2.0.
  • Версия : 0.0.0
  • Расколы :
Расколоть Примеры
'default' 59
'train' 43
'validation' 16
  • Функции :
{
    "idx": {
        "dtype": "int32",
        "id": null,
        "_type": "Value"
    },
    "inputs": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_scores": {
        "feature": {
            "dtype": "int32",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

формальные_fallacies_syllogisms_negation

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

ds = tfds.load('huggingface:bigbench/formal_fallacies_syllogisms_negation')
  • Описание :
The Beyond the Imitation Game Benchmark (BIG-bench) is a collaborative benchmark intended to
probe large language models, and extrapolate their future capabilities.
  • Лицензия : Лицензия Apache 2.0.
  • Версия : 0.0.0
  • Расколы :
Расколоть Примеры
'default' 14200
'train' 11360
'validation' 2840
  • Функции :
{
    "idx": {
        "dtype": "int32",
        "id": null,
        "_type": "Value"
    },
    "inputs": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_scores": {
        "feature": {
            "dtype": "int32",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

драгоценный камень

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

ds = tfds.load('huggingface:bigbench/gem')
  • Описание :
The Beyond the Imitation Game Benchmark (BIG-bench) is a collaborative benchmark intended to
probe large language models, and extrapolate their future capabilities.
  • Лицензия : Лицензия Apache 2.0.
  • Версия : 0.0.0
  • Расколы :
Расколоть Примеры
'default' 14802
'train' 11845
'validation' 2957
  • Функции :
{
    "idx": {
        "dtype": "int32",
        "id": null,
        "_type": "Value"
    },
    "inputs": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_scores": {
        "feature": {
            "dtype": "int32",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

гендер_inclusive_sentences_german

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

ds = tfds.load('huggingface:bigbench/gender_inclusive_sentences_german')
  • Описание :
The Beyond the Imitation Game Benchmark (BIG-bench) is a collaborative benchmark intended to
probe large language models, and extrapolate their future capabilities.
  • Лицензия : Лицензия Apache 2.0.
  • Версия : 0.0.0
  • Расколы :
Расколоть Примеры
'default' 200
'train' 160
'validation' 40
  • Функции :
{
    "idx": {
        "dtype": "int32",
        "id": null,
        "_type": "Value"
    },
    "inputs": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_scores": {
        "feature": {
            "dtype": "int32",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

общее_знание

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

ds = tfds.load('huggingface:bigbench/general_knowledge')
  • Описание :
The Beyond the Imitation Game Benchmark (BIG-bench) is a collaborative benchmark intended to
probe large language models, and extrapolate their future capabilities.
  • Лицензия : Лицензия Apache 2.0.
  • Версия : 0.0.0
  • Расколы :
Расколоть Примеры
'default' 70
'train' 54
'validation' 16
  • Функции :
{
    "idx": {
        "dtype": "int32",
        "id": null,
        "_type": "Value"
    },
    "inputs": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_scores": {
        "feature": {
            "dtype": "int32",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

геометрические_фигуры

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

ds = tfds.load('huggingface:bigbench/geometric_shapes')
  • Описание :
The Beyond the Imitation Game Benchmark (BIG-bench) is a collaborative benchmark intended to
probe large language models, and extrapolate their future capabilities.
  • Лицензия : Лицензия Apache 2.0.
  • Версия : 0.0.0
  • Расколы :
Расколоть Примеры
'default' 359
'train' 288
'validation' 71
  • Функции :
{
    "idx": {
        "dtype": "int32",
        "id": null,
        "_type": "Value"
    },
    "inputs": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_scores": {
        "feature": {
            "dtype": "int32",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

target_step_wikihow

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

ds = tfds.load('huggingface:bigbench/goal_step_wikihow')
  • Описание :
The Beyond the Imitation Game Benchmark (BIG-bench) is a collaborative benchmark intended to
probe large language models, and extrapolate their future capabilities.
  • Лицензия : Лицензия Apache 2.0.
  • Версия : 0.0.0
  • Расколы :
Расколоть Примеры
'default' 7053
'train' 5643
'validation' 1410
  • Функции :
{
    "idx": {
        "dtype": "int32",
        "id": null,
        "_type": "Value"
    },
    "inputs": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_scores": {
        "feature": {
            "dtype": "int32",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

gre_reading_comrehension

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

ds = tfds.load('huggingface:bigbench/gre_reading_comprehension')
  • Описание :
The Beyond the Imitation Game Benchmark (BIG-bench) is a collaborative benchmark intended to
probe large language models, and extrapolate their future capabilities.
  • Лицензия : Лицензия Apache 2.0.
  • Версия : 0.0.0
  • Расколы :
Расколоть Примеры
'default' 31
'train' 15
'validation' 16
  • Функции :
{
    "idx": {
        "dtype": "int32",
        "id": null,
        "_type": "Value"
    },
    "inputs": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_scores": {
        "feature": {
            "dtype": "int32",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

ххх_alignment

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

ds = tfds.load('huggingface:bigbench/hhh_alignment')
  • Описание :
The Beyond the Imitation Game Benchmark (BIG-bench) is a collaborative benchmark intended to
probe large language models, and extrapolate their future capabilities.
  • Лицензия : Лицензия Apache 2.0.
  • Версия : 0.0.0
  • Расколы :
Расколоть Примеры
'default' 221
'train' 179
'validation' 42
  • Функции :
{
    "idx": {
        "dtype": "int32",
        "id": null,
        "_type": "Value"
    },
    "inputs": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_scores": {
        "feature": {
            "dtype": "int32",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

хинди_вопрос_ответ

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

ds = tfds.load('huggingface:bigbench/hindi_question_answering')
  • Описание :
The Beyond the Imitation Game Benchmark (BIG-bench) is a collaborative benchmark intended to
probe large language models, and extrapolate their future capabilities.
  • Лицензия : Лицензия Apache 2.0.
  • Версия : 0.0.0
  • Расколы :
Расколоть Примеры
'default' 6610
'train' 5288
'validation' 1322
  • Функции :
{
    "idx": {
        "dtype": "int32",
        "id": null,
        "_type": "Value"
    },
    "inputs": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_scores": {
        "feature": {
            "dtype": "int32",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

индуистское_знание

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

ds = tfds.load('huggingface:bigbench/hindu_knowledge')
  • Описание :
The Beyond the Imitation Game Benchmark (BIG-bench) is a collaborative benchmark intended to
probe large language models, and extrapolate their future capabilities.
  • Лицензия : Лицензия Apache 2.0.
  • Версия : 0.0.0
  • Расколы :
Расколоть Примеры
'default' 175
'train' 140
'validation' 35
  • Функции :
{
    "idx": {
        "dtype": "int32",
        "id": null,
        "_type": "Value"
    },
    "inputs": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_scores": {
        "feature": {
            "dtype": "int32",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

хинглиш_токсичность

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

ds = tfds.load('huggingface:bigbench/hinglish_toxicity')
  • Описание :
The Beyond the Imitation Game Benchmark (BIG-bench) is a collaborative benchmark intended to
probe large language models, and extrapolate their future capabilities.
  • Лицензия : Лицензия Apache 2.0.
  • Версия : 0.0.0
  • Расколы :
Расколоть Примеры
'default' 200
'train' 160
'validation' 40
  • Функции :
{
    "idx": {
        "dtype": "int32",
        "id": null,
        "_type": "Value"
    },
    "inputs": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_scores": {
        "feature": {
            "dtype": "int32",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

human_organs_senses

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

ds = tfds.load('huggingface:bigbench/human_organs_senses')
  • Описание :
The Beyond the Imitation Game Benchmark (BIG-bench) is a collaborative benchmark intended to
probe large language models, and extrapolate their future capabilities.
  • Лицензия : Лицензия Apache 2.0.
  • Версия : 0.0.0
  • Расколы :
Расколоть Примеры
'default' 42
'train' 26
'validation' 16
  • Функции :
{
    "idx": {
        "dtype": "int32",
        "id": null,
        "_type": "Value"
    },
    "inputs": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_scores": {
        "feature": {
            "dtype": "int32",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

гипербатон

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

ds = tfds.load('huggingface:bigbench/hyperbaton')
  • Описание :
The Beyond the Imitation Game Benchmark (BIG-bench) is a collaborative benchmark intended to
probe large language models, and extrapolate their future capabilities.
  • Лицензия : Лицензия Apache 2.0.
  • Версия : 0.0.0
  • Расколы :
Расколоть Примеры
'default' 50000
'train' 40000
'validation' 10000
  • Функции :
{
    "idx": {
        "dtype": "int32",
        "id": null,
        "_type": "Value"
    },
    "inputs": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_scores": {
        "feature": {
            "dtype": "int32",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

идентификация_math_theorems

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

ds = tfds.load('huggingface:bigbench/identify_math_theorems')
  • Описание :
The Beyond the Imitation Game Benchmark (BIG-bench) is a collaborative benchmark intended to
probe large language models, and extrapolate their future capabilities.
  • Лицензия : Лицензия Apache 2.0.
  • Версия : 0.0.0
  • Расколы :
Расколоть Примеры
'default' 53
'train' 37
'validation' 16
  • Функции :
{
    "idx": {
        "dtype": "int32",
        "id": null,
        "_type": "Value"
    },
    "inputs": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_scores": {
        "feature": {
            "dtype": "int32",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

идентифицировать_нечетную_метафору

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

ds = tfds.load('huggingface:bigbench/identify_odd_metaphor')
  • Описание :
The Beyond the Imitation Game Benchmark (BIG-bench) is a collaborative benchmark intended to
probe large language models, and extrapolate their future capabilities.
  • Лицензия : Лицензия Apache 2.0.
  • Версия : 0.0.0
  • Расколы :
Расколоть Примеры
'default' 47
'train' 31
'validation' 16
  • Функции :
{
    "idx": {
        "dtype": "int32",
        "id": null,
        "_type": "Value"
    },
    "inputs": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_scores": {
        "feature": {
            "dtype": "int32",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

импликатуры

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

ds = tfds.load('huggingface:bigbench/implicatures')
  • Описание :
The Beyond the Imitation Game Benchmark (BIG-bench) is a collaborative benchmark intended to
probe large language models, and extrapolate their future capabilities.
  • Лицензия : Лицензия Apache 2.0.
  • Версия : 0.0.0
  • Расколы :
Расколоть Примеры
'default' 492
'train' 394
'validation' 98
  • Функции :
{
    "idx": {
        "dtype": "int32",
        "id": null,
        "_type": "Value"
    },
    "inputs": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_scores": {
        "feature": {
            "dtype": "int32",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

implicit_relations

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

ds = tfds.load('huggingface:bigbench/implicit_relations')
  • Описание :
The Beyond the Imitation Game Benchmark (BIG-bench) is a collaborative benchmark intended to
probe large language models, and extrapolate their future capabilities.
  • Лицензия : Лицензия Apache 2.0.
  • Версия : 0.0.0
  • Расколы :
Расколоть Примеры
'default' 85
'train' 68
'validation' 17
  • Функции :
{
    "idx": {
        "dtype": "int32",
        "id": null,
        "_type": "Value"
    },
    "inputs": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_scores": {
        "feature": {
            "dtype": "int32",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

намерение_распознавание

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

ds = tfds.load('huggingface:bigbench/intent_recognition')
  • Описание :
The Beyond the Imitation Game Benchmark (BIG-bench) is a collaborative benchmark intended to
probe large language models, and extrapolate their future capabilities.
  • Лицензия : Лицензия Apache 2.0.
  • Версия : 0.0.0
  • Расколы :
Расколоть Примеры
'default' 693
'train' 555
'validation' 138
  • Функции :
{
    "idx": {
        "dtype": "int32",
        "id": null,
        "_type": "Value"
    },
    "inputs": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_scores": {
        "feature": {
            "dtype": "int32",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

International_phonetic_alphabet_nli

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

ds = tfds.load('huggingface:bigbench/international_phonetic_alphabet_nli')
  • Описание :
The Beyond the Imitation Game Benchmark (BIG-bench) is a collaborative benchmark intended to
probe large language models, and extrapolate their future capabilities.
  • Лицензия : Лицензия Apache 2.0.
  • Версия : 0.0.0
  • Расколы :
Расколоть Примеры
'default' 126
'train' 101
'validation' 25
  • Функции :
{
    "idx": {
        "dtype": "int32",
        "id": null,
        "_type": "Value"
    },
    "inputs": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_scores": {
        "feature": {
            "dtype": "int32",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

международный_фонетический_алфавит_транслитерация

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

ds = tfds.load('huggingface:bigbench/international_phonetic_alphabet_transliterate')
  • Описание :
The Beyond the Imitation Game Benchmark (BIG-bench) is a collaborative benchmark intended to
probe large language models, and extrapolate their future capabilities.
  • Лицензия : Лицензия Apache 2.0.
  • Версия : 0.0.0
  • Расколы :
Расколоть Примеры
'default' 1003
'train' 803
'validation' 200
  • Функции :
{
    "idx": {
        "dtype": "int32",
        "id": null,
        "_type": "Value"
    },
    "inputs": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_scores": {
        "feature": {
            "dtype": "int32",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

intersect_geometry

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

ds = tfds.load('huggingface:bigbench/intersect_geometry')
  • Описание :
The Beyond the Imitation Game Benchmark (BIG-bench) is a collaborative benchmark intended to
probe large language models, and extrapolate their future capabilities.
  • Лицензия : Лицензия Apache 2.0.
  • Версия : 0.0.0
  • Расколы :
Расколоть Примеры
'default' 249999
'train' 200000
'validation' 49999
  • Функции :
{
    "idx": {
        "dtype": "int32",
        "id": null,
        "_type": "Value"
    },
    "inputs": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_scores": {
        "feature": {
            "dtype": "int32",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

ирония_идентификация

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

ds = tfds.load('huggingface:bigbench/irony_identification')
  • Описание :
The Beyond the Imitation Game Benchmark (BIG-bench) is a collaborative benchmark intended to
probe large language models, and extrapolate their future capabilities.
  • Лицензия : Лицензия Apache 2.0.
  • Версия : 0.0.0
  • Расколы :
Расколоть Примеры
'default' 99
'train' 80
'validation' 19
  • Функции :
{
    "idx": {
        "dtype": "int32",
        "id": null,
        "_type": "Value"
    },
    "inputs": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_scores": {
        "feature": {
            "dtype": "int32",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

kanji_ascii

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

ds = tfds.load('huggingface:bigbench/kanji_ascii')
  • Описание :
The Beyond the Imitation Game Benchmark (BIG-bench) is a collaborative benchmark intended to
probe large language models, and extrapolate their future capabilities.
  • Лицензия : Лицензия Apache 2.0.
  • Версия : 0.0.0
  • Расколы :
Расколоть Примеры
'default' 1092
'train' 875
'validation' 217
  • Функции :
{
    "idx": {
        "dtype": "int32",
        "id": null,
        "_type": "Value"
    },
    "inputs": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_scores": {
        "feature": {
            "dtype": "int32",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

каннада

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

ds = tfds.load('huggingface:bigbench/kannada')
  • Описание :
The Beyond the Imitation Game Benchmark (BIG-bench) is a collaborative benchmark intended to
probe large language models, and extrapolate their future capabilities.
  • Лицензия : Лицензия Apache 2.0.
  • Версия : 0.0.0
  • Расколы :
Расколоть Примеры
'default' 316
'train' 253
'validation' 63
  • Функции :
{
    "idx": {
        "dtype": "int32",
        "id": null,
        "_type": "Value"
    },
    "inputs": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_scores": {
        "feature": {
            "dtype": "int32",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

key_value_maps

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

ds = tfds.load('huggingface:bigbench/key_value_maps')
  • Описание :
The Beyond the Imitation Game Benchmark (BIG-bench) is a collaborative benchmark intended to
probe large language models, and extrapolate their future capabilities.
  • Лицензия : Лицензия Apache 2.0.
  • Версия : 0.0.0
  • Расколы :
Расколоть Примеры
'default' 101
'train' 80
'validation' 21
  • Функции :
{
    "idx": {
        "dtype": "int32",
        "id": null,
        "_type": "Value"
    },
    "inputs": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_scores": {
        "feature": {
            "dtype": "int32",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

известные_неизвестные

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

ds = tfds.load('huggingface:bigbench/known_unknowns')
  • Описание :
The Beyond the Imitation Game Benchmark (BIG-bench) is a collaborative benchmark intended to
probe large language models, and extrapolate their future capabilities.
  • Лицензия : Лицензия Apache 2.0.
  • Версия : 0.0.0
  • Расколы :
Расколоть Примеры
'default' 46
'train' 30
'validation' 16
  • Функции :
{
    "idx": {
        "dtype": "int32",
        "id": null,
        "_type": "Value"
    },
    "inputs": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_scores": {
        "feature": {
            "dtype": "int32",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

языковые_игры

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

ds = tfds.load('huggingface:bigbench/language_games')
  • Описание :
The Beyond the Imitation Game Benchmark (BIG-bench) is a collaborative benchmark intended to
probe large language models, and extrapolate their future capabilities.
  • Лицензия : Лицензия Apache 2.0.
  • Версия : 0.0.0
  • Расколы :
Расколоть Примеры
'default' 2128
'train' 1704 г.
'validation' 424
  • Функции :
{
    "idx": {
        "dtype": "int32",
        "id": null,
        "_type": "Value"
    },
    "inputs": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_scores": {
        "feature": {
            "dtype": "int32",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

языковая_идентификация

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

ds = tfds.load('huggingface:bigbench/language_identification')
  • Описание :
The Beyond the Imitation Game Benchmark (BIG-bench) is a collaborative benchmark intended to
probe large language models, and extrapolate their future capabilities.
  • Лицензия : Лицензия Apache 2.0.
  • Версия : 0.0.0
  • Расколы :
Расколоть Примеры
'default' 10000
'train' 8000
'validation' 2000 г.
  • Функции :
{
    "idx": {
        "dtype": "int32",
        "id": null,
        "_type": "Value"
    },
    "inputs": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_scores": {
        "feature": {
            "dtype": "int32",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

лингвистические_отображения

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

ds = tfds.load('huggingface:bigbench/linguistic_mappings')
  • Описание :
The Beyond the Imitation Game Benchmark (BIG-bench) is a collaborative benchmark intended to
probe large language models, and extrapolate their future capabilities.
  • Лицензия : Лицензия Apache 2.0.
  • Версия : 0.0.0
  • Расколы :
Расколоть Примеры
'default' 15527
'train' 12426
'validation' 3101
  • Функции :
{
    "idx": {
        "dtype": "int32",
        "id": null,
        "_type": "Value"
    },
    "inputs": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_scores": {
        "feature": {
            "dtype": "int32",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

лингвистика_пазлы

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

ds = tfds.load('huggingface:bigbench/linguistics_puzzles')
  • Описание :
The Beyond the Imitation Game Benchmark (BIG-bench) is a collaborative benchmark intended to
probe large language models, and extrapolate their future capabilities.
  • Лицензия : Лицензия Apache 2.0.
  • Версия : 0.0.0
  • Расколы :
Расколоть Примеры
'default' 2000 г.
'train' 1600
'validation' 400
  • Функции :
{
    "idx": {
        "dtype": "int32",
        "id": null,
        "_type": "Value"
    },
    "inputs": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_scores": {
        "feature": {
            "dtype": "int32",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

список_функций

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

ds = tfds.load('huggingface:bigbench/list_functions')
  • Описание :
The Beyond the Imitation Game Benchmark (BIG-bench) is a collaborative benchmark intended to
probe large language models, and extrapolate their future capabilities.
  • Лицензия : Лицензия Apache 2.0.
  • Версия : 0.0.0
  • Расколы :
Расколоть Примеры
'default' 10750
'train' 8700
'validation' 2050 год
  • Функции :
{
    "idx": {
        "dtype": "int32",
        "id": null,
        "_type": "Value"
    },
    "inputs": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_scores": {
        "feature": {
            "dtype": "int32",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

logic_grid_puzzle

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

ds = tfds.load('huggingface:bigbench/logic_grid_puzzle')
  • Описание :
The Beyond the Imitation Game Benchmark (BIG-bench) is a collaborative benchmark intended to
probe large language models, and extrapolate their future capabilities.
  • Лицензия : Лицензия Apache 2.0.
  • Версия : 0.0.0
  • Расколы :
Расколоть Примеры
'default' 1000
'train' 800
'validation' 200
  • Функции :
{
    "idx": {
        "dtype": "int32",
        "id": null,
        "_type": "Value"
    },
    "inputs": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_scores": {
        "feature": {
            "dtype": "int32",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

логические_арги

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

ds = tfds.load('huggingface:bigbench/logical_args')
  • Описание :
The Beyond the Imitation Game Benchmark (BIG-bench) is a collaborative benchmark intended to
probe large language models, and extrapolate their future capabilities.
  • Лицензия : Лицензия Apache 2.0.
  • Версия : 0.0.0
  • Расколы :
Расколоть Примеры
'default' 32
'train' 16
'validation' 16
  • Функции :
{
    "idx": {
        "dtype": "int32",
        "id": null,
        "_type": "Value"
    },
    "inputs": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_scores": {
        "feature": {
            "dtype": "int32",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

логический_вывод

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

ds = tfds.load('huggingface:bigbench/logical_deduction')
  • Описание :
The Beyond the Imitation Game Benchmark (BIG-bench) is a collaborative benchmark intended to
probe large language models, and extrapolate their future capabilities.
  • Лицензия : Лицензия Apache 2.0.
  • Версия : 0.0.0
  • Расколы :
Расколоть Примеры
'default' 1500
'train' 1200
'validation' 300
  • Функции :
{
    "idx": {
        "dtype": "int32",
        "id": null,
        "_type": "Value"
    },
    "inputs": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_scores": {
        "feature": {
            "dtype": "int32",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

логический_fallacy_detection

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

ds = tfds.load('huggingface:bigbench/logical_fallacy_detection')
  • Описание :
The Beyond the Imitation Game Benchmark (BIG-bench) is a collaborative benchmark intended to
probe large language models, and extrapolate their future capabilities.
  • Лицензия : Лицензия Apache 2.0.
  • Версия : 0.0.0
  • Расколы :
Расколоть Примеры
'default' 2800
'train' 2240
'validation' 560
  • Функции :
{
    "idx": {
        "dtype": "int32",
        "id": null,
        "_type": "Value"
    },
    "inputs": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_scores": {
        "feature": {
            "dtype": "int32",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

логическая_последовательность

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

ds = tfds.load('huggingface:bigbench/logical_sequence')
  • Описание :
The Beyond the Imitation Game Benchmark (BIG-bench) is a collaborative benchmark intended to
probe large language models, and extrapolate their future capabilities.
  • Лицензия : Лицензия Apache 2.0.
  • Версия : 0.0.0
  • Расколы :
Расколоть Примеры
'default' 39
'train' 23
'validation' 16
  • Функции :
{
    "idx": {
        "dtype": "int32",
        "id": null,
        "_type": "Value"
    },
    "inputs": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_scores": {
        "feature": {
            "dtype": "int32",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

математическая_индукция

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

ds = tfds.load('huggingface:bigbench/mathematical_induction')
  • Описание :
The Beyond the Imitation Game Benchmark (BIG-bench) is a collaborative benchmark intended to
probe large language models, and extrapolate their future capabilities.
  • Лицензия : Лицензия Apache 2.0.
  • Версия : 0.0.0
  • Расколы :
Расколоть Примеры
'default' 69
'train' 53
'validation' 16
  • Функции :
{
    "idx": {
        "dtype": "int32",
        "id": null,
        "_type": "Value"
    },
    "inputs": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_scores": {
        "feature": {
            "dtype": "int32",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

матрицыформы

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

ds = tfds.load('huggingface:bigbench/matrixshapes')
  • Описание :
The Beyond the Imitation Game Benchmark (BIG-bench) is a collaborative benchmark intended to
probe large language models, and extrapolate their future capabilities.
  • Лицензия : Лицензия Apache 2.0.
  • Версия : 0.0.0
  • Расколы :
Расколоть Примеры
'default' 4462
'train' 3570
'validation' 892
  • Функции :
{
    "idx": {
        "dtype": "int32",
        "id": null,
        "_type": "Value"
    },
    "inputs": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_scores": {
        "feature": {
            "dtype": "int32",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

метафора_логическое значение

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

ds = tfds.load('huggingface:bigbench/metaphor_boolean')
  • Описание :
The Beyond the Imitation Game Benchmark (BIG-bench) is a collaborative benchmark intended to
probe large language models, and extrapolate their future capabilities.
  • Лицензия : Лицензия Apache 2.0.
  • Версия : 0.0.0
  • Расколы :
Расколоть Примеры
'default' 680
'train' 544
'validation' 136
  • Функции :
{
    "idx": {
        "dtype": "int32",
        "id": null,
        "_type": "Value"
    },
    "inputs": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_scores": {
        "feature": {
            "dtype": "int32",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

метафора_понимание

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

ds = tfds.load('huggingface:bigbench/metaphor_understanding')
  • Описание :
The Beyond the Imitation Game Benchmark (BIG-bench) is a collaborative benchmark intended to
probe large language models, and extrapolate their future capabilities.
  • Лицензия : Лицензия Apache 2.0.
  • Версия : 0.0.0
  • Расколы :
Расколоть Примеры
'default' 234
'train' 188
'validation' 46
  • Функции :
{
    "idx": {
        "dtype": "int32",
        "id": null,
        "_type": "Value"
    },
    "inputs": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_scores": {
        "feature": {
            "dtype": "int32",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

минуту_mysteries_qa

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

ds = tfds.load('huggingface:bigbench/minute_mysteries_qa')
  • Описание :
The Beyond the Imitation Game Benchmark (BIG-bench) is a collaborative benchmark intended to
probe large language models, and extrapolate their future capabilities.
  • Лицензия : Лицензия Apache 2.0.
  • Версия : 0.0.0
  • Расколы :
Расколоть Примеры
'default' 477
'train' 383
'validation' 94
  • Функции :
{
    "idx": {
        "dtype": "int32",
        "id": null,
        "_type": "Value"
    },
    "inputs": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_scores": {
        "feature": {
            "dtype": "int32",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

заблуждения

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

ds = tfds.load('huggingface:bigbench/misconceptions')
  • Описание :
The Beyond the Imitation Game Benchmark (BIG-bench) is a collaborative benchmark intended to
probe large language models, and extrapolate their future capabilities.
  • Лицензия : Лицензия Apache 2.0.
  • Версия : 0.0.0
  • Расколы :
Расколоть Примеры
'default' 219
'train' 176
'validation' 43
  • Функции :
{
    "idx": {
        "dtype": "int32",
        "id": null,
        "_type": "Value"
    },
    "inputs": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_scores": {
        "feature": {
            "dtype": "int32",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

заблуждения_русский

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

ds = tfds.load('huggingface:bigbench/misconceptions_russian')
  • Описание :
The Beyond the Imitation Game Benchmark (BIG-bench) is a collaborative benchmark intended to
probe large language models, and extrapolate their future capabilities.
  • Лицензия : Лицензия Apache 2.0.
  • Версия : 0.0.0
  • Расколы :
Расколоть Примеры
'default' 49
'train' 33
'validation' 16
  • Функции :
{
    "idx": {
        "dtype": "int32",
        "id": null,
        "_type": "Value"
    },
    "inputs": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_scores": {
        "feature": {
            "dtype": "int32",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

mnist_ascii

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

ds = tfds.load('huggingface:bigbench/mnist_ascii')
  • Описание :
The Beyond the Imitation Game Benchmark (BIG-bench) is a collaborative benchmark intended to
probe large language models, and extrapolate their future capabilities.
  • Лицензия : Лицензия Apache 2.0.
  • Версия : 0.0.0
  • Расколы :
Расколоть Примеры
'default' 69984
'train' 55988
'validation' 13996
  • Функции :
{
    "idx": {
        "dtype": "int32",
        "id": null,
        "_type": "Value"
    },
    "inputs": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_scores": {
        "feature": {
            "dtype": "int32",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

модифицированная_арифметика

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

ds = tfds.load('huggingface:bigbench/modified_arithmetic')
  • Описание :
The Beyond the Imitation Game Benchmark (BIG-bench) is a collaborative benchmark intended to
probe large language models, and extrapolate their future capabilities.
  • Лицензия : Лицензия Apache 2.0.
  • Версия : 0.0.0
  • Расколы :
Расколоть Примеры
'default' 6000
'train' 4800
'validation' 1200
  • Функции :
{
    "idx": {
        "dtype": "int32",
        "id": null,
        "_type": "Value"
    },
    "inputs": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_scores": {
        "feature": {
            "dtype": "int32",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

моральная_допустимость

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

ds = tfds.load('huggingface:bigbench/moral_permissibility')
  • Описание :
The Beyond the Imitation Game Benchmark (BIG-bench) is a collaborative benchmark intended to
probe large language models, and extrapolate their future capabilities.
  • Лицензия : Лицензия Apache 2.0.
  • Версия : 0.0.0
  • Расколы :
Расколоть Примеры
'default' 342
'train' 274
'validation' 68
  • Функции :
{
    "idx": {
        "dtype": "int32",
        "id": null,
        "_type": "Value"
    },
    "inputs": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_scores": {
        "feature": {
            "dtype": "int32",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

фильм_диалог_тот же_или_различный

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

ds = tfds.load('huggingface:bigbench/movie_dialog_same_or_different')
  • Описание :
The Beyond the Imitation Game Benchmark (BIG-bench) is a collaborative benchmark intended to
probe large language models, and extrapolate their future capabilities.
  • Лицензия : Лицензия Apache 2.0.
  • Версия : 0.0.0
  • Расколы :
Расколоть Примеры
'default' 50000
'train' 40000
'validation' 10000
  • Функции :
{
    "idx": {
        "dtype": "int32",
        "id": null,
        "_type": "Value"
    },
    "inputs": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_scores": {
        "feature": {
            "dtype": "int32",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

фильм_рекомендация

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

ds = tfds.load('huggingface:bigbench/movie_recommendation')
  • Описание :
The Beyond the Imitation Game Benchmark (BIG-bench) is a collaborative benchmark intended to
probe large language models, and extrapolate their future capabilities.
  • Лицензия : Лицензия Apache 2.0.
  • Версия : 0.0.0
  • Расколы :
Расколоть Примеры
'default' 500
'train' 400
'validation' 100
  • Функции :
{
    "idx": {
        "dtype": "int32",
        "id": null,
        "_type": "Value"
    },
    "inputs": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_scores": {
        "feature": {
            "dtype": "int32",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

mult_data_wrangling

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

ds = tfds.load('huggingface:bigbench/mult_data_wrangling')
  • Описание :
The Beyond the Imitation Game Benchmark (BIG-bench) is a collaborative benchmark intended to
probe large language models, and extrapolate their future capabilities.
  • Лицензия : Лицензия Apache 2.0.
  • Версия : 0.0.0
  • Расколы :
Расколоть Примеры
'default' 7854
'train' 6380
'validation' 1474
  • Функции :
{
    "idx": {
        "dtype": "int32",
        "id": null,
        "_type": "Value"
    },
    "inputs": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_scores": {
        "feature": {
            "dtype": "int32",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

мультиэмо

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

ds = tfds.load('huggingface:bigbench/multiemo')
  • Описание :
The Beyond the Imitation Game Benchmark (BIG-bench) is a collaborative benchmark intended to
probe large language models, and extrapolate their future capabilities.
  • Лицензия : Лицензия Apache 2.0.
  • Версия : 0.0.0
  • Расколы :
Расколоть Примеры
'default' 1437281
'train' 1149873
'validation' 287408
  • Функции :
{
    "idx": {
        "dtype": "int32",
        "id": null,
        "_type": "Value"
    },
    "inputs": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_scores": {
        "feature": {
            "dtype": "int32",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

natural_instructions

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

ds = tfds.load('huggingface:bigbench/natural_instructions')
  • Описание :
The Beyond the Imitation Game Benchmark (BIG-bench) is a collaborative benchmark intended to
probe large language models, and extrapolate their future capabilities.
  • Лицензия : Лицензия Apache 2.0.
  • Версия : 0.0.0
  • Расколы :
Расколоть Примеры
'default' 193250
'train' 154615
'validation' 38635
  • Функции :
{
    "idx": {
        "dtype": "int32",
        "id": null,
        "_type": "Value"
    },
    "inputs": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_scores": {
        "feature": {
            "dtype": "int32",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

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

ds = tfds.load('huggingface:bigbench/navigate')
  • Описание :
The Beyond the Imitation Game Benchmark (BIG-bench) is a collaborative benchmark intended to
probe large language models, and extrapolate their future capabilities.
  • Лицензия : Лицензия Apache 2.0.
  • Версия : 0.0.0
  • Расколы :
Расколоть Примеры
'default' 1000
'train' 800
'validation' 200
  • Функции :
{
    "idx": {
        "dtype": "int32",
        "id": null,
        "_type": "Value"
    },
    "inputs": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_scores": {
        "feature": {
            "dtype": "int32",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

бессмысленные_слова_грамматика

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

ds = tfds.load('huggingface:bigbench/nonsense_words_grammar')
  • Описание :
The Beyond the Imitation Game Benchmark (BIG-bench) is a collaborative benchmark intended to
probe large language models, and extrapolate their future capabilities.
  • Лицензия : Лицензия Apache 2.0.
  • Версия : 0.0.0
  • Расколы :
Расколоть Примеры
'default' 50
'train' 34
'validation' 16
  • Функции :
{
    "idx": {
        "dtype": "int32",
        "id": null,
        "_type": "Value"
    },
    "inputs": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_scores": {
        "feature": {
            "dtype": "int32",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

роман_концепции

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

ds = tfds.load('huggingface:bigbench/novel_concepts')
  • Описание :
The Beyond the Imitation Game Benchmark (BIG-bench) is a collaborative benchmark intended to
probe large language models, and extrapolate their future capabilities.
  • Лицензия : Лицензия Apache 2.0.
  • Версия : 0.0.0
  • Расколы :
Расколоть Примеры
'default' 32
'train' 16
'validation' 16
  • Функции :
{
    "idx": {
        "dtype": "int32",
        "id": null,
        "_type": "Value"
    },
    "inputs": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_scores": {
        "feature": {
            "dtype": "int32",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

объект_counting

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

ds = tfds.load('huggingface:bigbench/object_counting')
  • Описание :
The Beyond the Imitation Game Benchmark (BIG-bench) is a collaborative benchmark intended to
probe large language models, and extrapolate their future capabilities.
  • Лицензия : Лицензия Apache 2.0.
  • Версия : 0.0.0
  • Расколы :
Расколоть Примеры
'default' 1000
'train' 800
'validation' 200
  • Функции :
{
    "idx": {
        "dtype": "int32",
        "id": null,
        "_type": "Value"
    },
    "inputs": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_scores": {
        "feature": {
            "dtype": "int32",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

нечетный_один_out

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

ds = tfds.load('huggingface:bigbench/odd_one_out')
  • Описание :
The Beyond the Imitation Game Benchmark (BIG-bench) is a collaborative benchmark intended to
probe large language models, and extrapolate their future capabilities.
  • Лицензия : Лицензия Apache 2.0.
  • Версия : 0.0.0
  • Расколы :
Расколоть Примеры
'default' 86
'train' 69
'validation' 17
  • Функции :
{
    "idx": {
        "dtype": "int32",
        "id": null,
        "_type": "Value"
    },
    "inputs": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_scores": {
        "feature": {
            "dtype": "int32",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

операторы

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

ds = tfds.load('huggingface:bigbench/operators')
  • Описание :
The Beyond the Imitation Game Benchmark (BIG-bench) is a collaborative benchmark intended to
probe large language models, and extrapolate their future capabilities.
  • Лицензия : Лицензия Apache 2.0.
  • Версия : 0.0.0
  • Расколы :
Расколоть Примеры
'default' 210
'train' 168
'validation' 42
  • Функции :
{
    "idx": {
        "dtype": "int32",
        "id": null,
        "_type": "Value"
    },
    "inputs": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_scores": {
        "feature": {
            "dtype": "int32",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

параграф_сегментация

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

ds = tfds.load('huggingface:bigbench/paragraph_segmentation')
  • Описание :
The Beyond the Imitation Game Benchmark (BIG-bench) is a collaborative benchmark intended to
probe large language models, and extrapolate their future capabilities.
  • Лицензия : Лицензия Apache 2.0.
  • Версия : 0.0.0
  • Расколы :
Расколоть Примеры
'default' 9000
'train' 7200
'validation' 1800 г.
  • Функции :
{
    "idx": {
        "dtype": "int32",
        "id": null,
        "_type": "Value"
    },
    "inputs": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_scores": {
        "feature": {
            "dtype": "int32",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

parsinlu_qa

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

ds = tfds.load('huggingface:bigbench/parsinlu_qa')
  • Описание :
The Beyond the Imitation Game Benchmark (BIG-bench) is a collaborative benchmark intended to
probe large language models, and extrapolate their future capabilities.
  • Лицензия : Лицензия Apache 2.0.
  • Версия : 0.0.0
  • Расколы :
Расколоть Примеры
'default' 1050
'train' 840
'validation' 210
  • Функции :
{
    "idx": {
        "dtype": "int32",
        "id": null,
        "_type": "Value"
    },
    "inputs": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_scores": {
        "feature": {
            "dtype": "int32",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

parsinlu_reading_comrehension

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

ds = tfds.load('huggingface:bigbench/parsinlu_reading_comprehension')
  • Описание :
The Beyond the Imitation Game Benchmark (BIG-bench) is a collaborative benchmark intended to
probe large language models, and extrapolate their future capabilities.
  • Лицензия : Лицензия Apache 2.0.
  • Версия : 0.0.0
  • Расколы :
Расколоть Примеры
'default' 518
'train' 415
'validation' 103
  • Функции :
{
    "idx": {
        "dtype": "int32",
        "id": null,
        "_type": "Value"
    },
    "inputs": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_scores": {
        "feature": {
            "dtype": "int32",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

penguins_in_a_table

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

ds = tfds.load('huggingface:bigbench/penguins_in_a_table')
  • Описание :
The Beyond the Imitation Game Benchmark (BIG-bench) is a collaborative benchmark intended to
probe large language models, and extrapolate their future capabilities.
  • Лицензия : Лицензия Apache 2.0.
  • Версия : 0.0.0
  • Расколы :
Расколоть Примеры
'default' 149
'train' 120
'validation' 29
  • Функции :
{
    "idx": {
        "dtype": "int32",
        "id": null,
        "_type": "Value"
    },
    "inputs": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_scores": {
        "feature": {
            "dtype": "int32",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

периодические_элементы

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

ds = tfds.load('huggingface:bigbench/periodic_elements')
  • Описание :
The Beyond the Imitation Game Benchmark (BIG-bench) is a collaborative benchmark intended to
probe large language models, and extrapolate their future capabilities.
  • Лицензия : Лицензия Apache 2.0.
  • Версия : 0.0.0
  • Расколы :
Расколоть Примеры
'default' 654
'train' 524
'validation' 130
  • Функции :
{
    "idx": {
        "dtype": "int32",
        "id": null,
        "_type": "Value"
    },
    "inputs": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_scores": {
        "feature": {
            "dtype": "int32",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

персидские_идиомы

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

ds = tfds.load('huggingface:bigbench/persian_idioms')
  • Описание :
The Beyond the Imitation Game Benchmark (BIG-bench) is a collaborative benchmark intended to
probe large language models, and extrapolate their future capabilities.
  • Лицензия : Лицензия Apache 2.0.
  • Версия : 0.0.0
  • Расколы :
Расколоть Примеры
'default' 66
'train' 50
'validation' 16
  • Функции :
{
    "idx": {
        "dtype": "int32",
        "id": null,
        "_type": "Value"
    },
    "inputs": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_scores": {
        "feature": {
            "dtype": "int32",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

фраза_родственность

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

ds = tfds.load('huggingface:bigbench/phrase_relatedness')
  • Описание :
The Beyond the Imitation Game Benchmark (BIG-bench) is a collaborative benchmark intended to
probe large language models, and extrapolate their future capabilities.
  • Лицензия : Лицензия Apache 2.0.
  • Версия : 0.0.0
  • Расколы :
Расколоть Примеры
'default' 100
'train' 80
'validation' 20
  • Функции :
{
    "idx": {
        "dtype": "int32",
        "id": null,
        "_type": "Value"
    },
    "inputs": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_scores": {
        "feature": {
            "dtype": "int32",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

физическая_интуиция

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

ds = tfds.load('huggingface:bigbench/physical_intuition')
  • Описание :
The Beyond the Imitation Game Benchmark (BIG-bench) is a collaborative benchmark intended to
probe large language models, and extrapolate their future capabilities.
  • Лицензия : Лицензия Apache 2.0.
  • Версия : 0.0.0
  • Расколы :
Расколоть Примеры
'default' 81
'train' 65
'validation' 16
  • Функции :
{
    "idx": {
        "dtype": "int32",
        "id": null,
        "_type": "Value"
    },
    "inputs": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_scores": {
        "feature": {
            "dtype": "int32",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

физика

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

ds = tfds.load('huggingface:bigbench/physics')
  • Описание :
The Beyond the Imitation Game Benchmark (BIG-bench) is a collaborative benchmark intended to
probe large language models, and extrapolate their future capabilities.
  • Лицензия : Лицензия Apache 2.0.
  • Версия : 0.0.0
  • Расколы :
Расколоть Примеры
'default' 229
'train' 184
'validation' 45
  • Функции :
{
    "idx": {
        "dtype": "int32",
        "id": null,
        "_type": "Value"
    },
    "inputs": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_scores": {
        "feature": {
            "dtype": "int32",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

физика_вопросы

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

ds = tfds.load('huggingface:bigbench/physics_questions')
  • Описание :
The Beyond the Imitation Game Benchmark (BIG-bench) is a collaborative benchmark intended to
probe large language models, and extrapolate their future capabilities.
  • Лицензия : Лицензия Apache 2.0.
  • Версия : 0.0.0
  • Расколы :
Расколоть Примеры
'default' 54
'train' 38
'validation' 16
  • Функции :
{
    "idx": {
        "dtype": "int32",
        "id": null,
        "_type": "Value"
    },
    "inputs": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_scores": {
        "feature": {
            "dtype": "int32",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

play_dialog_same_or_различные

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

ds = tfds.load('huggingface:bigbench/play_dialog_same_or_different')
  • Описание :
The Beyond the Imitation Game Benchmark (BIG-bench) is a collaborative benchmark intended to
probe large language models, and extrapolate their future capabilities.
  • Лицензия : Лицензия Apache 2.0.
  • Версия : 0.0.0
  • Расколы :
Расколоть Примеры
'default' 3264
'train' 2612
'validation' 652
  • Функции :
{
    "idx": {
        "dtype": "int32",
        "id": null,
        "_type": "Value"
    },
    "inputs": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_scores": {
        "feature": {
            "dtype": "int32",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

Polish_sequence_labeling

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

ds = tfds.load('huggingface:bigbench/polish_sequence_labeling')
  • Описание :
The Beyond the Imitation Game Benchmark (BIG-bench) is a collaborative benchmark intended to
probe large language models, and extrapolate their future capabilities.
  • Лицензия : Лицензия Apache 2.0.
  • Версия : 0.0.0
  • Расколы :
Расколоть Примеры
'default' 12812
'train' 10250
'validation' 2562
  • Функции :
{
    "idx": {
        "dtype": "int32",
        "id": null,
        "_type": "Value"
    },
    "inputs": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_scores": {
        "feature": {
            "dtype": "int32",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

пресуппозиции_as_nli

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

ds = tfds.load('huggingface:bigbench/presuppositions_as_nli')
  • Описание :
The Beyond the Imitation Game Benchmark (BIG-bench) is a collaborative benchmark intended to
probe large language models, and extrapolate their future capabilities.
  • Лицензия : Лицензия Apache 2.0.
  • Версия : 0.0.0
  • Расколы :
Расколоть Примеры
'default' 735
'train' 588
'validation' 147
  • Функции :
{
    "idx": {
        "dtype": "int32",
        "id": null,
        "_type": "Value"
    },
    "inputs": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_scores": {
        "feature": {
            "dtype": "int32",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

qa_wikidata

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

ds = tfds.load('huggingface:bigbench/qa_wikidata')
  • Описание :
The Beyond the Imitation Game Benchmark (BIG-bench) is a collaborative benchmark intended to
probe large language models, and extrapolate their future capabilities.
  • Лицензия : Лицензия Apache 2.0.
  • Версия : 0.0.0
  • Расколы :
Расколоть Примеры
'default' 20321
'train' 16257
'validation' 4064
  • Функции :
{
    "idx": {
        "dtype": "int32",
        "id": null,
        "_type": "Value"
    },
    "inputs": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_scores": {
        "feature": {
            "dtype": "int32",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

вопрос_выбор

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

ds = tfds.load('huggingface:bigbench/question_selection')
  • Описание :
The Beyond the Imitation Game Benchmark (BIG-bench) is a collaborative benchmark intended to
probe large language models, and extrapolate their future capabilities.
  • Лицензия : Лицензия Apache 2.0.
  • Версия : 0.0.0
  • Расколы :
Расколоть Примеры
'default' 1582 г.
'train' 1266
'validation' 316
  • Функции :
{
    "idx": {
        "dtype": "int32",
        "id": null,
        "_type": "Value"
    },
    "inputs": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_scores": {
        "feature": {
            "dtype": "int32",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

реальный_или_фальшивый_текст

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

ds = tfds.load('huggingface:bigbench/real_or_fake_text')
  • Описание :
The Beyond the Imitation Game Benchmark (BIG-bench) is a collaborative benchmark intended to
probe large language models, and extrapolate their future capabilities.
  • Лицензия : Лицензия Apache 2.0.
  • Версия : 0.0.0
  • Расколы :
Расколоть Примеры
'default' 15088
'train' 12072
'validation' 3016
  • Функции :
{
    "idx": {
        "dtype": "int32",
        "id": null,
        "_type": "Value"
    },
    "inputs": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_scores": {
        "feature": {
            "dtype": "int32",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

рассуждения_о_цветных_объектах

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

ds = tfds.load('huggingface:bigbench/reasoning_about_colored_objects')
  • Описание :
The Beyond the Imitation Game Benchmark (BIG-bench) is a collaborative benchmark intended to
probe large language models, and extrapolate their future capabilities.
  • Лицензия : Лицензия Apache 2.0.
  • Версия : 0.0.0
  • Расколы :
Расколоть Примеры
'default' 2000 г.
'train' 1600
'validation' 400
  • Функции :
{
    "idx": {
        "dtype": "int32",
        "id": null,
        "_type": "Value"
    },
    "inputs": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_scores": {
        "feature": {
            "dtype": "int32",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

повтор_копия_логика

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

ds = tfds.load('huggingface:bigbench/repeat_copy_logic')
  • Описание :
The Beyond the Imitation Game Benchmark (BIG-bench) is a collaborative benchmark intended to
probe large language models, and extrapolate their future capabilities.
  • Лицензия : Лицензия Apache 2.0.
  • Версия : 0.0.0
  • Расколы :
Расколоть Примеры
'default' 32
'train' 16
'validation' 16
  • Функции :
{
    "idx": {
        "dtype": "int32",
        "id": null,
        "_type": "Value"
    },
    "inputs": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_scores": {
        "feature": {
            "dtype": "int32",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

перефразировать

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

ds = tfds.load('huggingface:bigbench/rephrase')
  • Описание :
The Beyond the Imitation Game Benchmark (BIG-bench) is a collaborative benchmark intended to
probe large language models, and extrapolate their future capabilities.
  • Лицензия : Лицензия Apache 2.0.
  • Версия : 0.0.0
  • Расколы :
Расколоть Примеры
'default' 78
'train' 62
'validation' 16
  • Функции :
{
    "idx": {
        "dtype": "int32",
        "id": null,
        "_type": "Value"
    },
    "inputs": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_scores": {
        "feature": {
            "dtype": "int32",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

загадка_чувство

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

ds = tfds.load('huggingface:bigbench/riddle_sense')
  • Описание :
The Beyond the Imitation Game Benchmark (BIG-bench) is a collaborative benchmark intended to
probe large language models, and extrapolate their future capabilities.
  • Лицензия : Лицензия Apache 2.0.
  • Версия : 0.0.0
  • Расколы :
Расколоть Примеры
'default' 49
'train' 33
'validation' 16
  • Функции :
{
    "idx": {
        "dtype": "int32",
        "id": null,
        "_type": "Value"
    },
    "inputs": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_scores": {
        "feature": {
            "dtype": "int32",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

руины_имена

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

ds = tfds.load('huggingface:bigbench/ruin_names')
  • Описание :
The Beyond the Imitation Game Benchmark (BIG-bench) is a collaborative benchmark intended to
probe large language models, and extrapolate their future capabilities.
  • Лицензия : Лицензия Apache 2.0.
  • Версия : 0.0.0
  • Расколы :
Расколоть Примеры
'default' 448
'train' 359
'validation' 89
  • Функции :
{
    "idx": {
        "dtype": "int32",
        "id": null,
        "_type": "Value"
    },
    "inputs": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_scores": {
        "feature": {
            "dtype": "int32",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

salient_translation_error_detection

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

ds = tfds.load('huggingface:bigbench/salient_translation_error_detection')
  • Описание :
The Beyond the Imitation Game Benchmark (BIG-bench) is a collaborative benchmark intended to
probe large language models, and extrapolate their future capabilities.
  • Лицензия : Лицензия Apache 2.0.
  • Версия : 0.0.0
  • Расколы :
Расколоть Примеры
'default' 998
'train' 799
'validation' 199
  • Функции :
{
    "idx": {
        "dtype": "int32",
        "id": null,
        "_type": "Value"
    },
    "inputs": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_scores": {
        "feature": {
            "dtype": "int32",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

научный_пресс_релиз

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

ds = tfds.load('huggingface:bigbench/scientific_press_release')
  • Описание :
The Beyond the Imitation Game Benchmark (BIG-bench) is a collaborative benchmark intended to
probe large language models, and extrapolate their future capabilities.
  • Лицензия : Лицензия Apache 2.0.
  • Версия : 0.0.0
  • Расколы :
Расколоть Примеры
'default' 50
'train' 34
'validation' 16
  • Функции :
{
    "idx": {
        "dtype": "int32",
        "id": null,
        "_type": "Value"
    },
    "inputs": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_scores": {
        "feature": {
            "dtype": "int32",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

semantic_parsing_in_context_sparc

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

ds = tfds.load('huggingface:bigbench/semantic_parsing_in_context_sparc')
  • Описание :
The Beyond the Imitation Game Benchmark (BIG-bench) is a collaborative benchmark intended to
probe large language models, and extrapolate their future capabilities.
  • Лицензия : Лицензия Apache 2.0.
  • Версия : 0.0.0
  • Расколы :
Расколоть Примеры
'default' 1155
'train' 924
'validation' 231
  • Функции :
{
    "idx": {
        "dtype": "int32",
        "id": null,
        "_type": "Value"
    },
    "inputs": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_scores": {
        "feature": {
            "dtype": "int32",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

semantic_parsing_spider

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

ds = tfds.load('huggingface:bigbench/semantic_parsing_spider')
  • Описание :
The Beyond the Imitation Game Benchmark (BIG-bench) is a collaborative benchmark intended to
probe large language models, and extrapolate their future capabilities.
  • Лицензия : Лицензия Apache 2.0.
  • Версия : 0.0.0
  • Расколы :
Расколоть Примеры
'default' 1034
'train' 828
'validation' 206
  • Функции :
{
    "idx": {
        "dtype": "int32",
        "id": null,
        "_type": "Value"
    },
    "inputs": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_scores": {
        "feature": {
            "dtype": "int32",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

предложение_двусмысленность

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

ds = tfds.load('huggingface:bigbench/sentence_ambiguity')
  • Описание :
The Beyond the Imitation Game Benchmark (BIG-bench) is a collaborative benchmark intended to
probe large language models, and extrapolate their future capabilities.
  • Лицензия : Лицензия Apache 2.0.
  • Версия : 0.0.0
  • Расколы :
Расколоть Примеры
'default' 60
'train' 44
'validation' 16
  • Функции :
{
    "idx": {
        "dtype": "int32",
        "id": null,
        "_type": "Value"
    },
    "inputs": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_scores": {
        "feature": {
            "dtype": "int32",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

сходства_абстракция

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

ds = tfds.load('huggingface:bigbench/similarities_abstraction')
  • Описание :
The Beyond the Imitation Game Benchmark (BIG-bench) is a collaborative benchmark intended to
probe large language models, and extrapolate their future capabilities.
  • Лицензия : Лицензия Apache 2.0.
  • Версия : 0.0.0
  • Расколы :
Расколоть Примеры
'default' 76
'train' 60
'validation' 16
  • Функции :
{
    "idx": {
        "dtype": "int32",
        "id": null,
        "_type": "Value"
    },
    "inputs": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_scores": {
        "feature": {
            "dtype": "int32",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

simp_turing_concept

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

ds = tfds.load('huggingface:bigbench/simp_turing_concept')
  • Описание :
The Beyond the Imitation Game Benchmark (BIG-bench) is a collaborative benchmark intended to
probe large language models, and extrapolate their future capabilities.
  • Лицензия : Лицензия Apache 2.0.
  • Версия : 0.0.0
  • Расколы :
Расколоть Примеры
'default' 6390
'train' 5112
'validation' 1278
  • Функции :
{
    "idx": {
        "dtype": "int32",
        "id": null,
        "_type": "Value"
    },
    "inputs": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_scores": {
        "feature": {
            "dtype": "int32",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

simple_arithmetic_json

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

ds = tfds.load('huggingface:bigbench/simple_arithmetic_json')
  • Описание :
The Beyond the Imitation Game Benchmark (BIG-bench) is a collaborative benchmark intended to
probe large language models, and extrapolate their future capabilities.
  • Лицензия : Лицензия Apache 2.0.
  • Версия : 0.0.0
  • Расколы :
Расколоть Примеры
'default' 30
'train' 14
'validation' 16
  • Функции :
{
    "idx": {
        "dtype": "int32",
        "id": null,
        "_type": "Value"
    },
    "inputs": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_scores": {
        "feature": {
            "dtype": "int32",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

simple_arithmetic_json_multiple_choice

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

ds = tfds.load('huggingface:bigbench/simple_arithmetic_json_multiple_choice')
  • Описание :
The Beyond the Imitation Game Benchmark (BIG-bench) is a collaborative benchmark intended to
probe large language models, and extrapolate their future capabilities.
  • Лицензия : Лицензия Apache 2.0.
  • Версия : 0.0.0
  • Расколы :
Расколоть Примеры
'default' 8
'train' 0
'validation' 0
  • Функции :
{
    "idx": {
        "dtype": "int32",
        "id": null,
        "_type": "Value"
    },
    "inputs": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_scores": {
        "feature": {
            "dtype": "int32",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

simple_arithmetic_json_subtasks

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

ds = tfds.load('huggingface:bigbench/simple_arithmetic_json_subtasks')
  • Описание :
The Beyond the Imitation Game Benchmark (BIG-bench) is a collaborative benchmark intended to
probe large language models, and extrapolate their future capabilities.
  • Лицензия : Лицензия Apache 2.0.
  • Версия : 0.0.0
  • Расколы :
Расколоть Примеры
'default' 30
'train' 15
'validation' 15
  • Функции :
{
    "idx": {
        "dtype": "int32",
        "id": null,
        "_type": "Value"
    },
    "inputs": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_scores": {
        "feature": {
            "dtype": "int32",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

simple_arithmetic_multiple_targets_json

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

ds = tfds.load('huggingface:bigbench/simple_arithmetic_multiple_targets_json')
  • Описание :
The Beyond the Imitation Game Benchmark (BIG-bench) is a collaborative benchmark intended to
probe large language models, and extrapolate their future capabilities.
  • Лицензия : Лицензия Apache 2.0.
  • Версия : 0.0.0
  • Расколы :
Расколоть Примеры
'default' 10
'train' 0
'validation' 0
  • Функции :
{
    "idx": {
        "dtype": "int32",
        "id": null,
        "_type": "Value"
    },
    "inputs": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_scores": {
        "feature": {
            "dtype": "int32",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

simple_ethical_questions

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

ds = tfds.load('huggingface:bigbench/simple_ethical_questions')
  • Описание :
The Beyond the Imitation Game Benchmark (BIG-bench) is a collaborative benchmark intended to
probe large language models, and extrapolate their future capabilities.
  • Лицензия : Лицензия Apache 2.0.
  • Версия : 0.0.0
  • Расколы :
Расколоть Примеры
'default' 115
'train' 92
'validation' 23
  • Функции :
{
    "idx": {
        "dtype": "int32",
        "id": null,
        "_type": "Value"
    },
    "inputs": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_scores": {
        "feature": {
            "dtype": "int32",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

simple_text_editing

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

ds = tfds.load('huggingface:bigbench/simple_text_editing')
  • Описание :
The Beyond the Imitation Game Benchmark (BIG-bench) is a collaborative benchmark intended to
probe large language models, and extrapolate their future capabilities.
  • Лицензия : Лицензия Apache 2.0.
  • Версия : 0.0.0
  • Расколы :
Расколоть Примеры
'default' 47
'train' 31
'validation' 16
  • Функции :
{
    "idx": {
        "dtype": "int32",
        "id": null,
        "_type": "Value"
    },
    "inputs": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_scores": {
        "feature": {
            "dtype": "int32",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

язвит

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

ds = tfds.load('huggingface:bigbench/snarks')
  • Описание :
The Beyond the Imitation Game Benchmark (BIG-bench) is a collaborative benchmark intended to
probe large language models, and extrapolate their future capabilities.
  • Лицензия : Лицензия Apache 2.0.
  • Версия : 0.0.0
  • Расколы :
Расколоть Примеры
'default' 181
'train' 145
'validation' 36
  • Функции :
{
    "idx": {
        "dtype": "int32",
        "id": null,
        "_type": "Value"
    },
    "inputs": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_scores": {
        "feature": {
            "dtype": "int32",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

Social_iqa

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

ds = tfds.load('huggingface:bigbench/social_iqa')
  • Описание :
The Beyond the Imitation Game Benchmark (BIG-bench) is a collaborative benchmark intended to
probe large language models, and extrapolate their future capabilities.
  • Лицензия : Лицензия Apache 2.0.
  • Версия : 0.0.0
  • Расколы :
Расколоть Примеры
'default' 1935 год
'train' 1548 г.
'validation' 387
  • Функции :
{
    "idx": {
        "dtype": "int32",
        "id": null,
        "_type": "Value"
    },
    "inputs": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_scores": {
        "feature": {
            "dtype": "int32",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

социальная_поддержка

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

ds = tfds.load('huggingface:bigbench/social_support')
  • Описание :
The Beyond the Imitation Game Benchmark (BIG-bench) is a collaborative benchmark intended to
probe large language models, and extrapolate their future capabilities.
  • Лицензия : Лицензия Apache 2.0.
  • Версия : 0.0.0
  • Расколы :
Расколоть Примеры
'default' 897
'train' 718
'validation' 179
  • Функции :
{
    "idx": {
        "dtype": "int32",
        "id": null,
        "_type": "Value"
    },
    "inputs": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_scores": {
        "feature": {
            "dtype": "int32",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

спортивное_понимание

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

ds = tfds.load('huggingface:bigbench/sports_understanding')
  • Описание :
The Beyond the Imitation Game Benchmark (BIG-bench) is a collaborative benchmark intended to
probe large language models, and extrapolate their future capabilities.
  • Лицензия : Лицензия Apache 2.0.
  • Версия : 0.0.0
  • Расколы :
Расколоть Примеры
'default' 986
'train' 789
'validation' 197
  • Функции :
{
    "idx": {
        "dtype": "int32",
        "id": null,
        "_type": "Value"
    },
    "inputs": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_scores": {
        "feature": {
            "dtype": "int32",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

странные_истории

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

ds = tfds.load('huggingface:bigbench/strange_stories')
  • Описание :
The Beyond the Imitation Game Benchmark (BIG-bench) is a collaborative benchmark intended to
probe large language models, and extrapolate their future capabilities.
  • Лицензия : Лицензия Apache 2.0.
  • Версия : 0.0.0
  • Расколы :
Расколоть Примеры
'default' 174
'train' 140
'validation' 34
  • Функции :
{
    "idx": {
        "dtype": "int32",
        "id": null,
        "_type": "Value"
    },
    "inputs": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_scores": {
        "feature": {
            "dtype": "int32",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

стратегияqa

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

ds = tfds.load('huggingface:bigbench/strategyqa')
  • Описание :
The Beyond the Imitation Game Benchmark (BIG-bench) is a collaborative benchmark intended to
probe large language models, and extrapolate their future capabilities.
  • Лицензия : Лицензия Apache 2.0.
  • Версия : 0.0.0
  • Расколы :
Расколоть Примеры
'default' 2289
'train' 1832 г.
'validation' 457
  • Функции :
{
    "idx": {
        "dtype": "int32",
        "id": null,
        "_type": "Value"
    },
    "inputs": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_scores": {
        "feature": {
            "dtype": "int32",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

достаточная_информация

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

ds = tfds.load('huggingface:bigbench/sufficient_information')
  • Описание :
The Beyond the Imitation Game Benchmark (BIG-bench) is a collaborative benchmark intended to
probe large language models, and extrapolate their future capabilities.
  • Лицензия : Лицензия Apache 2.0.
  • Версия : 0.0.0
  • Расколы :
Расколоть Примеры
'default' 39
'train' 23
'validation' 16
  • Функции :
{
    "idx": {
        "dtype": "int32",
        "id": null,
        "_type": "Value"
    },
    "inputs": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_scores": {
        "feature": {
            "dtype": "int32",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

суицид_риск

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

ds = tfds.load('huggingface:bigbench/suicide_risk')
  • Описание :
The Beyond the Imitation Game Benchmark (BIG-bench) is a collaborative benchmark intended to
probe large language models, and extrapolate their future capabilities.
  • Лицензия : Лицензия Apache 2.0.
  • Версия : 0.0.0
  • Расколы :
Расколоть Примеры
'default' 40
'train' 24
'validation' 16
  • Функции :
{
    "idx": {
        "dtype": "int32",
        "id": null,
        "_type": "Value"
    },
    "inputs": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_scores": {
        "feature": {
            "dtype": "int32",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

суахили_английские_пословицы

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

ds = tfds.load('huggingface:bigbench/swahili_english_proverbs')
  • Описание :
The Beyond the Imitation Game Benchmark (BIG-bench) is a collaborative benchmark intended to
probe large language models, and extrapolate their future capabilities.
  • Лицензия : Лицензия Apache 2.0.
  • Версия : 0.0.0
  • Расколы :
Расколоть Примеры
'default' 153
'train' 123
'validation' 30
  • Функции :
{
    "idx": {
        "dtype": "int32",
        "id": null,
        "_type": "Value"
    },
    "inputs": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_scores": {
        "feature": {
            "dtype": "int32",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

шведские_то_немецкие_пословицы

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

ds = tfds.load('huggingface:bigbench/swedish_to_german_proverbs')
  • Описание :
The Beyond the Imitation Game Benchmark (BIG-bench) is a collaborative benchmark intended to
probe large language models, and extrapolate their future capabilities.
  • Лицензия : Лицензия Apache 2.0.
  • Версия : 0.0.0
  • Расколы :
Расколоть Примеры
'default' 72
'train' 56
'validation' 16
  • Функции :
{
    "idx": {
        "dtype": "int32",
        "id": null,
        "_type": "Value"
    },
    "inputs": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_scores": {
        "feature": {
            "dtype": "int32",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

символ_интерпретация

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

ds = tfds.load('huggingface:bigbench/symbol_interpretation')
  • Описание :
The Beyond the Imitation Game Benchmark (BIG-bench) is a collaborative benchmark intended to
probe large language models, and extrapolate their future capabilities.
  • Лицензия : Лицензия Apache 2.0.
  • Версия : 0.0.0
  • Расколы :
Расколоть Примеры
'default' 990
'train' 795
'validation' 195
  • Функции :
{
    "idx": {
        "dtype": "int32",
        "id": null,
        "_type": "Value"
    },
    "inputs": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_scores": {
        "feature": {
            "dtype": "int32",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

temporal_sequences

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

ds = tfds.load('huggingface:bigbench/temporal_sequences')
  • Описание :
The Beyond the Imitation Game Benchmark (BIG-bench) is a collaborative benchmark intended to
probe large language models, and extrapolate their future capabilities.
  • Лицензия : Лицензия Apache 2.0.
  • Версия : 0.0.0
  • Расколы :
Расколоть Примеры
'default' 1000
'train' 800
'validation' 200
  • Функции :
{
    "idx": {
        "dtype": "int32",
        "id": null,
        "_type": "Value"
    },
    "inputs": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_scores": {
        "feature": {
            "dtype": "int32",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

напряженный

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

ds = tfds.load('huggingface:bigbench/tense')
  • Описание :
The Beyond the Imitation Game Benchmark (BIG-bench) is a collaborative benchmark intended to
probe large language models, and extrapolate their future capabilities.
  • Лицензия : Лицензия Apache 2.0.
  • Версия : 0.0.0
  • Расколы :
Расколоть Примеры
'default' 286
'train' 229
'validation' 57
  • Функции :
{
    "idx": {
        "dtype": "int32",
        "id": null,
        "_type": "Value"
    },
    "inputs": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_scores": {
        "feature": {
            "dtype": "int32",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

хронометр

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

ds = tfds.load('huggingface:bigbench/timedial')
  • Описание :
The Beyond the Imitation Game Benchmark (BIG-bench) is a collaborative benchmark intended to
probe large language models, and extrapolate their future capabilities.
  • Лицензия : Лицензия Apache 2.0.
  • Версия : 0.0.0
  • Расколы :
Расколоть Примеры
'default' 2550
'train' 2040 год
'validation' 510
  • Функции :
{
    "idx": {
        "dtype": "int32",
        "id": null,
        "_type": "Value"
    },
    "inputs": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_scores": {
        "feature": {
            "dtype": "int32",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

тематический_чат

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

ds = tfds.load('huggingface:bigbench/topical_chat')
  • Описание :
The Beyond the Imitation Game Benchmark (BIG-bench) is a collaborative benchmark intended to
probe large language models, and extrapolate their future capabilities.
  • Лицензия : Лицензия Apache 2.0.
  • Версия : 0.0.0
  • Расколы :
Расколоть Примеры
'default' 22295
'train' 17836
'validation' 4459
  • Функции :
{
    "idx": {
        "dtype": "int32",
        "id": null,
        "_type": "Value"
    },
    "inputs": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_scores": {
        "feature": {
            "dtype": "int32",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

track_shuffled_objects

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

ds = tfds.load('huggingface:bigbench/tracking_shuffled_objects')
  • Описание :
The Beyond the Imitation Game Benchmark (BIG-bench) is a collaborative benchmark intended to
probe large language models, and extrapolate their future capabilities.
  • Лицензия : Лицензия Apache 2.0.
  • Версия : 0.0.0
  • Расколы :
Расколоть Примеры
'default' 3750
'train' 3000
'validation' 750
  • Функции :
{
    "idx": {
        "dtype": "int32",
        "id": null,
        "_type": "Value"
    },
    "inputs": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_scores": {
        "feature": {
            "dtype": "int32",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

понимание_басен

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

ds = tfds.load('huggingface:bigbench/understanding_fables')
  • Описание :
The Beyond the Imitation Game Benchmark (BIG-bench) is a collaborative benchmark intended to
probe large language models, and extrapolate their future capabilities.
  • Лицензия : Лицензия Apache 2.0.
  • Версия : 0.0.0
  • Расколы :
Расколоть Примеры
'default' 189
'train' 152
'validation' 37
  • Функции :
{
    "idx": {
        "dtype": "int32",
        "id": null,
        "_type": "Value"
    },
    "inputs": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_scores": {
        "feature": {
            "dtype": "int32",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

отменить_пермутацию

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

ds = tfds.load('huggingface:bigbench/undo_permutation')
  • Описание :
The Beyond the Imitation Game Benchmark (BIG-bench) is a collaborative benchmark intended to
probe large language models, and extrapolate their future capabilities.
  • Лицензия : Лицензия Apache 2.0.
  • Версия : 0.0.0
  • Расколы :
Расколоть Примеры
'default' 300
'train' 240
'validation' 60
  • Функции :
{
    "idx": {
        "dtype": "int32",
        "id": null,
        "_type": "Value"
    },
    "inputs": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_scores": {
        "feature": {
            "dtype": "int32",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

unit_conversion

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

ds = tfds.load('huggingface:bigbench/unit_conversion')
  • Описание :
The Beyond the Imitation Game Benchmark (BIG-bench) is a collaborative benchmark intended to
probe large language models, and extrapolate their future capabilities.
  • Лицензия : Лицензия Apache 2.0.
  • Версия : 0.0.0
  • Расколы :
Расколоть Примеры
'default' 23936
'train' 19151
'validation' 4785
  • Функции :
{
    "idx": {
        "dtype": "int32",
        "id": null,
        "_type": "Value"
    },
    "inputs": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_scores": {
        "feature": {
            "dtype": "int32",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

unit_interpretation

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

ds = tfds.load('huggingface:bigbench/unit_interpretation')
  • Описание :
The Beyond the Imitation Game Benchmark (BIG-bench) is a collaborative benchmark intended to
probe large language models, and extrapolate their future capabilities.
  • Лицензия : Лицензия Apache 2.0.
  • Версия : 0.0.0
  • Расколы :
Расколоть Примеры
'default' 100
'train' 80
'validation' 20
  • Функции :
{
    "idx": {
        "dtype": "int32",
        "id": null,
        "_type": "Value"
    },
    "inputs": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_scores": {
        "feature": {
            "dtype": "int32",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

unnatural_in_context_learning

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

ds = tfds.load('huggingface:bigbench/unnatural_in_context_learning')
  • Описание :
The Beyond the Imitation Game Benchmark (BIG-bench) is a collaborative benchmark intended to
probe large language models, and extrapolate their future capabilities.
  • Лицензия : Лицензия Apache 2.0.
  • Версия : 0.0.0
  • Расколы :
Расколоть Примеры
'default' 73420
'train' 58736
'validation' 14684
  • Функции :
{
    "idx": {
        "dtype": "int32",
        "id": null,
        "_type": "Value"
    },
    "inputs": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_scores": {
        "feature": {
            "dtype": "int32",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

витаминc_fact_verification

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

ds = tfds.load('huggingface:bigbench/vitaminc_fact_verification')
  • Описание :
The Beyond the Imitation Game Benchmark (BIG-bench) is a collaborative benchmark intended to
probe large language models, and extrapolate their future capabilities.
  • Лицензия : Лицензия Apache 2.0.
  • Версия : 0.0.0
  • Расколы :
Расколоть Примеры
'default' 54668
'train' 43735
'validation' 10933
  • Функции :
{
    "idx": {
        "dtype": "int32",
        "id": null,
        "_type": "Value"
    },
    "inputs": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_scores": {
        "feature": {
            "dtype": "int32",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

что_есть_тао

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

ds = tfds.load('huggingface:bigbench/what_is_the_tao')
  • Описание :
The Beyond the Imitation Game Benchmark (BIG-bench) is a collaborative benchmark intended to
probe large language models, and extrapolate their future capabilities.
  • Лицензия : Лицензия Apache 2.0.
  • Версия : 0.0.0
  • Расколы :
Расколоть Примеры
'default' 36
'train' 20
'validation' 16
  • Функции :
{
    "idx": {
        "dtype": "int32",
        "id": null,
        "_type": "Value"
    },
    "inputs": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_scores": {
        "feature": {
            "dtype": "int32",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

that_wiki_edit

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

ds = tfds.load('huggingface:bigbench/which_wiki_edit')
  • Описание :
The Beyond the Imitation Game Benchmark (BIG-bench) is a collaborative benchmark intended to
probe large language models, and extrapolate their future capabilities.
  • Лицензия : Лицензия Apache 2.0.
  • Версия : 0.0.0
  • Расколы :
Расколоть Примеры
'default' 571
'train' 457
'validation' 114
  • Функции :
{
    "idx": {
        "dtype": "int32",
        "id": null,
        "_type": "Value"
    },
    "inputs": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_scores": {
        "feature": {
            "dtype": "int32",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

Виновно, почему

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

ds = tfds.load('huggingface:bigbench/winowhy')
  • Описание :
The Beyond the Imitation Game Benchmark (BIG-bench) is a collaborative benchmark intended to
probe large language models, and extrapolate their future capabilities.
  • Лицензия : Лицензия Apache 2.0.
  • Версия : 0.0.0
  • Расколы :
Расколоть Примеры
'default' 2862
'train' 2290
'validation' 572
  • Функции :
{
    "idx": {
        "dtype": "int32",
        "id": null,
        "_type": "Value"
    },
    "inputs": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_scores": {
        "feature": {
            "dtype": "int32",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

сортировка слов

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

ds = tfds.load('huggingface:bigbench/word_sorting')
  • Описание :
The Beyond the Imitation Game Benchmark (BIG-bench) is a collaborative benchmark intended to
probe large language models, and extrapolate their future capabilities.
  • Лицензия : Лицензия Apache 2.0.
  • Версия : 0.0.0
  • Расколы :
Расколоть Примеры
'default' 1900 г.
'train' 1520 г.
'validation' 380
  • Функции :
{
    "idx": {
        "dtype": "int32",
        "id": null,
        "_type": "Value"
    },
    "inputs": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_scores": {
        "feature": {
            "dtype": "int32",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

word_unscrambling

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

ds = tfds.load('huggingface:bigbench/word_unscrambling')
  • Описание :
The Beyond the Imitation Game Benchmark (BIG-bench) is a collaborative benchmark intended to
probe large language models, and extrapolate their future capabilities.
  • Лицензия : Лицензия Apache 2.0.
  • Версия : 0.0.0
  • Расколы :
Расколоть Примеры
'default' 8917
'train' 7134
'validation' 1783 г.
  • Функции :
{
    "idx": {
        "dtype": "int32",
        "id": null,
        "_type": "Value"
    },
    "inputs": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_targets": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "multiple_choice_scores": {
        "feature": {
            "dtype": "int32",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}