Hugging Face's logo

Datasets:
shunk031
/
STAIR-Captions

Tasks:
Image-to-Text
Modalities:
Image
Text
Formats:
parquet
Sub-tasks:
image-captioning
Languages:
Japanese
Size:
100K - 1M
ArXiv:
Tags:
image-captioning
mscoco
Libraries:
Datasets
Dask
License:
Dataset card Viewer Files Community
1
STAIR-Captions / README.md
shunk031's picture
shunk031
Upload dataset
20bee3d verified
preview code
|
raw
history blame
4.14 kB
metadata
dataset_info:
  - config_name: v1.0.0
    features:
      - name: image_id
        dtype: int64
      - name: image
        dtype: image
      - name: file_name
        dtype: string
      - name: coco_url
        dtype: string
      - name: height
        dtype: int32
      - name: width
        dtype: int32
      - name: date_captured
        dtype: string
      - name: flickr_url
        dtype: string
      - name: license_id
        dtype: int32
      - name: license
        struct:
          - name: url
            dtype: string
          - name: license_id
            dtype: int8
          - name: name
            dtype: string
      - name: annotations
        sequence:
          - name: annotation_id
            dtype: int64
          - name: image_id
            dtype: int64
          - name: caption
            dtype: string
    splits:
      - name: train
        num_bytes: 13636500472.001
        num_examples: 82783
      - name: validation
        num_bytes: 6856928794.552
        num_examples: 40504
    download_size: 20132168975
    dataset_size: 20493429266.553
  - config_name: v1.1.0
    features:
      - name: image_id
        dtype: int64
      - name: image
        dtype: image
      - name: file_name
        dtype: string
      - name: coco_url
        dtype: string
      - name: height
        dtype: int32
      - name: width
        dtype: int32
      - name: date_captured
        dtype: string
      - name: flickr_url
        dtype: string
      - name: license_id
        dtype: int32
      - name: license
        struct:
          - name: url
            dtype: string
          - name: license_id
            dtype: int8
          - name: name
            dtype: string
      - name: annotations
        sequence:
          - name: annotation_id
            dtype: int64
          - name: image_id
            dtype: int64
          - name: caption
            dtype: string
    splits:
      - name: train
        num_bytes: 13636500472.001
        num_examples: 82783
      - name: validation
        num_bytes: 6856928794.552
        num_examples: 40504
    download_size: 20132168975
    dataset_size: 20493429266.553
  - config_name: v1.2.0
    features:
      - name: image_id
        dtype: int64
      - name: image
        dtype: image
      - name: file_name
        dtype: string
      - name: coco_url
        dtype: string
      - name: height
        dtype: int32
      - name: width
        dtype: int32
      - name: date_captured
        dtype: string
      - name: flickr_url
        dtype: string
      - name: license_id
        dtype: int32
      - name: license
        struct:
          - name: url
            dtype: string
          - name: license_id
            dtype: int8
          - name: name
            dtype: string
      - name: annotations
        sequence:
          - name: annotation_id
            dtype: int64
          - name: image_id
            dtype: int64
          - name: caption
            dtype: string
    splits:
      - name: train
        num_bytes: 13635411411.001
        num_examples: 82783
      - name: validation
        num_bytes: 6856395884.552
        num_examples: 40504
    download_size: 20131797179
    dataset_size: 20491807295.553
  - config_name: v1.2.0-tokenized
    features:
      - name: image_id
        dtype: int64
      - name: image
        dtype: image
      - name: file_name
        dtype: string
      - name: coco_url
        dtype: string
      - name: height
        dtype: int32
      - name: width
        dtype: int32
      - name: date_captured
        dtype: string
      - name: flickr_url
        dtype: string
      - name: license_id
        dtype: int32
      - name: license
        struct:
          - name: url
            dtype: string
          - name: license_id
            dtype: int8
          - name: name
            dtype: string
      - name: annotations
        sequence:
          - name: annotation_id
            dtype: int64
          - name: image_id
            dtype: int64
          - name: caption
            dtype: string
    splits:
      - name: train
        num_bytes: 13635411411.001
        num_examples: 82783
      - name: validation
        num_bytes: 6856395884.552
        num_examples: 40504
    download_size: 20131797179
    dataset_size: 20491807295.553
configs:
  - config_name: v1.0.0
    data_files:
      - split: train
        path: v1.0.0/train-*
      - split: validation
        path: v1.0.0/validation-*
  - config_name: v1.1.0
    data_files:
      - split: train
        path: v1.1.0/train-*
      - split: validation
        path: v1.1.0/validation-*
  - config_name: v1.2.0
    data_files:
      - split: train
        path: v1.2.0/train-*
      - split: validation
        path: v1.2.0/validation-*
  - config_name: v1.2.0-tokenized
    data_files:
      - split: train
        path: v1.2.0-tokenized/train-*
      - split: validation
        path: v1.2.0-tokenized/validation-*