концептуальные_подписи

Использованная литература:

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

ds = tfds.load('huggingface:conceptual_captions')
  • Описание :
Image captioning dataset
The resulting dataset (version 1.1) has been split into Training, Validation, and Test splits. The Training split consists of 3,318,333 image-URL/caption pairs, with a total number of 51,201 total token types in the captions (i.e., total vocabulary). The average number of tokens per captions is 10.3 (standard deviation of 4.5), while the median is 9.0 tokens per caption. The Validation split consists of 15,840 image-URL/caption pairs, with similar statistics.
  • Лицензия : Нет известной лицензии.
  • Версия : 1.1.0
  • Расколы :
Расколоть Примеры
'train' 3318333
'validation' 15840
  • Функции :
{
    "id": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "caption": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "url": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

немаркированный

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

ds = tfds.load('huggingface:conceptual_captions/unlabeled')
  • Описание :
Google's Conceptual Captions dataset has more than 3 million images, paired with natural-language captions.
In contrast with the curated style of the MS-COCO images, Conceptual Captions images and their raw descriptions are harvested from the web,
and therefore represent a wider variety of styles. The raw descriptions are harvested from the Alt-text HTML attribute associated with web images.
The authors developed an automatic pipeline that extracts, filters, and transforms candidate image/caption pairs, with the goal of achieving a balance of cleanliness,
informativeness, fluency, and learnability of the resulting captions.
  • Лицензия : Набор данных можно свободно использовать для любых целей, однако приветствуется упоминание Google LLC («Google») в качестве источника данных. Набор данных предоставляется «КАК ЕСТЬ» без каких-либо гарантий, явных или подразумеваемых. Google отказывается от любой ответственности за любой ущерб, прямой или косвенный, возникший в результате использования набора данных.

  • Версия : 0.0.0

  • Расколы :

Расколоть Примеры
'train' 3318333
'validation' 15840
  • Функции :
{
    "image_url": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "caption": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

помеченный

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

ds = tfds.load('huggingface:conceptual_captions/labeled')
  • Описание :
Google's Conceptual Captions dataset has more than 3 million images, paired with natural-language captions.
In contrast with the curated style of the MS-COCO images, Conceptual Captions images and their raw descriptions are harvested from the web,
and therefore represent a wider variety of styles. The raw descriptions are harvested from the Alt-text HTML attribute associated with web images.
The authors developed an automatic pipeline that extracts, filters, and transforms candidate image/caption pairs, with the goal of achieving a balance of cleanliness,
informativeness, fluency, and learnability of the resulting captions.
  • Лицензия : Набор данных можно свободно использовать для любых целей, однако приветствуется упоминание Google LLC («Google») в качестве источника данных. Набор данных предоставляется «КАК ЕСТЬ» без каких-либо гарантий, явных или подразумеваемых. Google отказывается от любой ответственности за любой ущерб, прямой или косвенный, возникший в результате использования набора данных.

  • Версия : 0.0.0

  • Расколы :

Расколоть Примеры
'train' 2007090
  • Функции :
{
    "image_url": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "caption": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "labels": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "MIDs": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "confidence_scores": {
        "feature": {
            "dtype": "float64",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}