sbu_captions

ข้อมูลอ้างอิง:

ใช้คำสั่งต่อไปนี้เพื่อโหลดชุดข้อมูลนี้ใน TFDS:

ds = tfds.load('huggingface:sbu_captions')
  • คำอธิบาย :
The SBU Captioned Photo Dataset is a collection of over 1 million images with associated text descriptions extracted from Flicker.
  • ใบอนุญาต : ไม่ทราบ
  • เวอร์ชัน : 0.0.0
  • แยก :
แยก ตัวอย่าง
'train' 1000000
  • คุณสมบัติ :
{
    "image_url": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "user_id": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "caption": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}