TwinDoc/funsd

---
dataset_info:
features:
- name: id
dtype: string
- name: words
sequence: string
- name: bboxes
sequence:
sequence: int64
- name: ner_tags
sequence:
class_label:
names:
'0': O
'1': B-HEADER
'2': I-HEADER
'3': B-QUESTION
'4': I-QUESTION
'5': B-ANSWER
'6': I-ANSWER
- name: image_path
dtype: string
- name: image
dtype: image
- name: bbox_id
sequence: int64
- name: words_sort_tlbr
sequence: string
- name: bboxes_sort_tlbr
sequence:
sequence: int64
- name: ner_tags_sort_tlbr
sequence: int64
- name: bbox_id__sort_tlbr
sequence: int64
splits:
- name: train
num_bytes: 14725225.0
num_examples: 149
- name: test
num_bytes: 5358339.0
num_examples: 50
- name: test_10
num_bytes: 1071658.2
num_examples: 10
download_size: 18253145
dataset_size: 21155222.2
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
- split: test
path: data/test-*
- split: test_10
path: data/test_10-*
---

原始信息汇总

数据集概述

数据集特征

• id: 字符串类型
• words: 字符串序列类型
• bboxes: 整数序列序列类型
• ner_tags: 类别标签序列类型，标签名包括：
  • 0: O
  • 1: B-HEADER
  • 2: I-HEADER
  • 3: B-QUESTION
  • 4: I-QUESTION
  • 5: B-ANSWER
  • 6: I-ANSWER
• image_path: 字符串类型
• image: 图像类型
• bbox_id: 整数序列类型
• words_sort_tlbr: 字符串序列类型
• bboxes_sort_tlbr: 整数序列序列类型
• ner_tags_sort_tlbr: 整数序列类型
• bbox_id__sort_tlbr: 整数序列类型

数据集分割

• train: 149个样本，占用14725225字节
• test: 50个样本，占用5358339字节
• test_10: 10个样本，占用1071658.2字节

数据集大小

• 下载大小: 18253145字节
• 数据集总大小: 21155222.2字节

配置文件

• default 配置:
  • 训练数据路径: data/train-*
  • 测试数据路径: data/test-*
  • 测试数据10路径: data/test_10-*