thaisum

อ้างอิง:

ไทยซัม

ใช้คำสั่งต่อไปนี้เพื่อโหลดชุดข้อมูลนี้ใน TFDS:

ds = tfds.load('huggingface:thaisum/thaisum')
  • คำอธิบาย :
ThaiSum is a large-scale corpus for Thai text summarization obtained from several online news websites namely Thairath,
ThaiPBS, Prachathai, and The Standard. This dataset consists of over 350,000 article and summary pairs
written by journalists.
  • ใบอนุญาต : ไม่มีใบอนุญาตที่รู้จัก
  • เวอร์ชัน : 1.0.0
  • แยก :
แยก ตัวอย่าง
'test' 11000
'train' 358868
'validation' 11000
  • คุณสมบัติ :
{
    "title": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "body": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "summary": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "type": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "tags": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "url": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}