md5 stringlengths 32 32 | image unknown | text stringlengths 1 179 | scene stringclasses 4
values | __index_level_0__ int64 0 4.19k |
|---|---|---|---|---|
930ac2240100e424b3fd46e799c85c5a | [
255,
216,
255,
224,
0,
16,
74,
70,
73,
70,
0,
1,
1,
0,
0,
1,
0,
1,
0,
0,
255,
219,
0,
67,
0,
8,
6,
6,
7,
6,
5,
8,
7,
7,
7,
9,
9,
8,
10,
12,
20,
13,
12,
11,
11,
12,
25,
18,
19,
15,
20,
29,
26,
31,
30,
29,
... | 高于全国农村的平均增长水平。三 | text_simplified_chinese | 0 |
ba9d1de550df380b51ceda2bf5a30457 | [
255,
216,
255,
224,
0,
16,
74,
70,
73,
70,
0,
1,
1,
0,
0,
1,
0,
1,
0,
0,
255,
219,
0,
67,
0,
8,
6,
6,
7,
6,
5,
8,
7,
7,
7,
9,
9,
8,
10,
12,
20,
13,
12,
11,
11,
12,
25,
18,
19,
15,
20,
29,
26,
31,
30,
29,
... | 据孙雷介绍，“十一五”以来，上海实施一系列强农惠农政 | text_simplified_chinese | 1 |
4d3c7a0561ee2915953ff2dd6f9af3d6 | [
255,
216,
255,
224,
0,
16,
74,
70,
73,
70,
0,
1,
1,
0,
0,
1,
0,
1,
0,
0,
255,
219,
0,
67,
0,
8,
6,
6,
7,
6,
5,
8,
7,
7,
7,
9,
9,
8,
10,
12,
20,
13,
12,
11,
11,
12,
25,
18,
19,
15,
20,
29,
26,
31,
30,
29,
... | 的 | text_simplified_chinese | 2 |
99ecf92be341b66d5a3eb95597a12ada | [
255,
216,
255,
224,
0,
16,
74,
70,
73,
70,
0,
1,
1,
0,
0,
1,
0,
1,
0,
0,
255,
219,
0,
67,
0,
8,
6,
6,
7,
6,
5,
8,
7,
7,
7,
9,
9,
8,
10,
12,
20,
13,
12,
11,
11,
12,
25,
18,
19,
15,
20,
29,
26,
31,
30,
29,
... | ·截至2023Q3全国国资背景基金备案数量累计9196只,基金规模累计8.91万亿元。基金注册区域集中于广东省、浙江省和江苏省,广东 | text_simplified_chinese | 3 |
e021f7594102eae34e2f59580ac385dd | [
255,
216,
255,
224,
0,
16,
74,
70,
73,
70,
0,
1,
1,
0,
0,
1,
0,
1,
0,
0,
255,
219,
0,
67,
0,
8,
6,
6,
7,
6,
5,
8,
7,
7,
7,
9,
9,
8,
10,
12,
20,
13,
12,
11,
11,
12,
25,
18,
19,
15,
20,
29,
26,
31,
30,
29,
... | 卢拉总统的政策，致力于消除贫 | text_simplified_chinese | 4 |
75823174e8bff3c5910c95bb6941ac27 | [
255,
216,
255,
224,
0,
16,
74,
70,
73,
70,
0,
1,
1,
0,
0,
1,
0,
1,
0,
0,
255,
219,
0,
67,
0,
8,
6,
6,
7,
6,
5,
8,
7,
7,
7,
9,
9,
8,
10,
12,
20,
13,
12,
11,
11,
12,
25,
18,
19,
15,
20,
29,
26,
31,
30,
29,
... | 村不同情况,灵活采用互换、转 | text_simplified_chinese | 5 |
5fc90c59abcd645f14ee11209ebf7ce5 | [
255,
216,
255,
224,
0,
16,
74,
70,
73,
70,
0,
1,
1,
0,
0,
1,
0,
1,
0,
0,
255,
219,
0,
67,
0,
8,
6,
6,
7,
6,
5,
8,
7,
7,
7,
9,
9,
8,
10,
12,
20,
13,
12,
11,
11,
12,
25,
18,
19,
15,
20,
29,
26,
31,
30,
29,
... | 融机构借贷:若成员有钱不还贷,保险公司就停止 | text_simplified_chinese | 6 |
ef4a537ca07a4a125c9ce330d0b3f41b | [
255,
216,
255,
224,
0,
16,
74,
70,
73,
70,
0,
1,
1,
0,
0,
1,
0,
1,
0,
0,
255,
219,
0,
67,
0,
8,
6,
6,
7,
6,
5,
8,
7,
7,
7,
9,
9,
8,
10,
12,
20,
13,
12,
11,
11,
12,
25,
18,
19,
15,
20,
29,
26,
31,
30,
29,
... | （小锣） | text_simplified_chinese | 7 |
305d51ac3aa2f2e0271ddbbe4b7bbf3a | [
255,
216,
255,
224,
0,
16,
74,
70,
73,
70,
0,
1,
1,
0,
0,
1,
0,
1,
0,
0,
255,
219,
0,
67,
0,
8,
6,
6,
7,
6,
5,
8,
7,
7,
7,
9,
9,
8,
10,
12,
20,
13,
12,
11,
11,
12,
25,
18,
19,
15,
20,
29,
26,
31,
30,
29,
... | 业。 | text_simplified_chinese | 8 |
19142cd56bab6d07a56894d9f0989e64 | [
255,
216,
255,
224,
0,
16,
74,
70,
73,
70,
0,
1,
1,
0,
0,
1,
0,
1,
0,
0,
255,
219,
0,
67,
0,
8,
6,
6,
7,
6,
5,
8,
7,
7,
7,
9,
9,
8,
10,
12,
20,
13,
12,
11,
11,
12,
25,
18,
19,
15,
20,
29,
26,
31,
30,
29,
... | 股票投资评级说明： | text_simplified_chinese | 9 |
7391e0d5c4f7b06a2f7dc70a01e12acd | [
255,
216,
255,
224,
0,
16,
74,
70,
73,
70,
0,
1,
1,
0,
0,
1,
0,
1,
0,
0,
255,
219,
0,
67,
0,
8,
6,
6,
7,
6,
5,
8,
7,
7,
7,
9,
9,
8,
10,
12,
20,
13,
12,
11,
11,
12,
25,
18,
19,
15,
20,
29,
26,
31,
30,
29,
... | 场（户）自身防御意识和应对自然灾害的能力。四 | text_simplified_chinese | 10 |
a4a99d3b833220b22990ad4d8ffe3444 | [
255,
216,
255,
224,
0,
16,
74,
70,
73,
70,
0,
1,
1,
0,
0,
1,
0,
1,
0,
0,
255,
219,
0,
67,
0,
8,
6,
6,
7,
6,
5,
8,
7,
7,
7,
9,
9,
8,
10,
12,
20,
13,
12,
11,
11,
12,
25,
18,
19,
15,
20,
29,
26,
31,
30,
29,
... | 社、中央各单位各部门出版社在内 | text_simplified_chinese | 11 |
2160b1e1700030ef8e16322cae19c5df | [
255,
216,
255,
224,
0,
16,
74,
70,
73,
70,
0,
1,
1,
0,
0,
1,
0,
1,
0,
0,
255,
219,
0,
67,
0,
8,
6,
6,
7,
6,
5,
8,
7,
7,
7,
9,
9,
8,
10,
12,
20,
13,
12,
11,
11,
12,
25,
18,
19,
15,
20,
29,
26,
31,
30,
29,
... | 除氢核外，原子核 | text_simplified_chinese | 12 |
f9809860a329a8cda89ea9b1099ef9f3 | [
255,
216,
255,
224,
0,
16,
74,
70,
73,
70,
0,
1,
1,
0,
0,
1,
0,
1,
0,
0,
255,
219,
0,
67,
0,
8,
6,
6,
7,
6,
5,
8,
7,
7,
7,
9,
9,
8,
10,
12,
20,
13,
12,
11,
11,
12,
25,
18,
19,
15,
20,
29,
26,
31,
30,
29,
... | (一)论述类文本阅读(本题共3小题,9分) | text_simplified_chinese | 13 |
d722e649c1b6e79bcde4bef2dff1ef6c | [
255,
216,
255,
224,
0,
16,
74,
70,
73,
70,
0,
1,
1,
0,
0,
1,
0,
1,
0,
0,
255,
219,
0,
67,
0,
8,
6,
6,
7,
6,
5,
8,
7,
7,
7,
9,
9,
8,
10,
12,
20,
13,
12,
11,
11,
12,
25,
18,
19,
15,
20,
29,
26,
31,
30,
29,
... | 管理的漏洞。 | text_simplified_chinese | 14 |
c818682ea1ae6151dbc32fc97fb7dac6 | [
255,
216,
255,
224,
0,
16,
74,
70,
73,
70,
0,
1,
1,
0,
0,
1,
0,
1,
0,
0,
255,
219,
0,
67,
0,
8,
6,
6,
7,
6,
5,
8,
7,
7,
7,
9,
9,
8,
10,
12,
20,
13,
12,
11,
11,
12,
25,
18,
19,
15,
20,
29,
26,
31,
30,
29,
... | 执政水平,为全面建设小康社会、坚 | text_simplified_chinese | 15 |
f1c6835d7acfe36901ebd4b94b05df34 | [
255,
216,
255,
224,
0,
16,
74,
70,
73,
70,
0,
1,
1,
0,
0,
1,
0,
1,
0,
0,
255,
219,
0,
67,
0,
8,
6,
6,
7,
6,
5,
8,
7,
7,
7,
9,
9,
8,
10,
12,
20,
13,
12,
11,
11,
12,
25,
18,
19,
15,
20,
29,
26,
31,
30,
29,
... | 人们注意到,5年里,中央政治局,中央政治局常 | text_simplified_chinese | 16 |
c53238d10c027e4b2bacf4fc30a70d6d | [
255,
216,
255,
224,
0,
16,
74,
70,
73,
70,
0,
1,
1,
0,
0,
1,
0,
1,
0,
0,
255,
219,
0,
67,
0,
8,
6,
6,
7,
6,
5,
8,
7,
7,
7,
9,
9,
8,
10,
12,
20,
13,
12,
11,
11,
12,
25,
18,
19,
15,
20,
29,
26,
31,
30,
29,
... | 习惯，不爱与你分享。这也许就是大多数法国人给 | text_simplified_chinese | 17 |
6bcea2d8956a9db364755a8b8db0babe | [
255,
216,
255,
224,
0,
16,
74,
70,
73,
70,
0,
1,
1,
0,
0,
1,
0,
1,
0,
0,
255,
219,
0,
67,
0,
8,
6,
6,
7,
6,
5,
8,
7,
7,
7,
9,
9,
8,
10,
12,
20,
13,
12,
11,
11,
12,
25,
18,
19,
15,
20,
29,
26,
31,
30,
29,
... | 产品加工水平和农畜产品品牌建设实现三个 | text_simplified_chinese | 18 |
f3b886111ef90747998d4f0325ca54da | [
255,
216,
255,
224,
0,
16,
74,
70,
73,
70,
0,
1,
1,
0,
0,
1,
0,
1,
0,
0,
255,
219,
0,
67,
0,
8,
6,
6,
7,
6,
5,
8,
7,
7,
7,
9,
9,
8,
10,
12,
20,
13,
12,
11,
11,
12,
25,
18,
19,
15,
20,
29,
26,
31,
30,
29,
... | 元，同时在报纸上发表致歉声明。 | text_simplified_chinese | 19 |
80fb47997c77ea343b131a5958da1282 | [
255,
216,
255,
224,
0,
16,
74,
70,
73,
70,
0,
1,
1,
0,
0,
1,
0,
1,
0,
0,
255,
219,
0,
67,
0,
8,
6,
6,
7,
6,
5,
8,
7,
7,
7,
9,
9,
8,
10,
12,
20,
13,
12,
11,
11,
12,
25,
18,
19,
15,
20,
29,
26,
31,
30,
29,
... | 民喜欢看、用得上、买得起的文化产品。要开 | text_simplified_chinese | 20 |
9a7246fb880f857d39189ad67a3075f3 | [
255,
216,
255,
224,
0,
16,
74,
70,
73,
70,
0,
1,
1,
0,
0,
1,
0,
1,
0,
0,
255,
219,
0,
67,
0,
8,
6,
6,
7,
6,
5,
8,
7,
7,
7,
9,
9,
8,
10,
12,
20,
13,
12,
11,
11,
12,
25,
18,
19,
15,
20,
29,
26,
31,
30,
29,
... | 二：输出的不仅是知识和价值观，更输出了无数参与祖国未来建设的栋梁 | text_simplified_chinese | 21 |
79835d8b3841275f37a433e5a70c2876 | [
255,
216,
255,
224,
0,
16,
74,
70,
73,
70,
0,
1,
1,
0,
0,
1,
0,
1,
0,
0,
255,
219,
0,
67,
0,
8,
6,
6,
7,
6,
5,
8,
7,
7,
7,
9,
9,
8,
10,
12,
20,
13,
12,
11,
11,
12,
25,
18,
19,
15,
20,
29,
26,
31,
30,
29,
... | 资金，实行以奖代补 | text_simplified_chinese | 22 |
f1343058f972cce2214596ea952b0e65 | [
255,
216,
255,
224,
0,
16,
74,
70,
73,
70,
0,
1,
1,
0,
0,
1,
0,
1,
0,
0,
255,
219,
0,
67,
0,
8,
6,
6,
7,
6,
5,
8,
7,
7,
7,
9,
9,
8,
10,
12,
20,
13,
12,
11,
11,
12,
25,
18,
19,
15,
20,
29,
26,
31,
30,
29,
... | 增加。2009年,福建全省农民人均收入为66802 | text_simplified_chinese | 23 |
fcc552faa19b4e7479e85d578e2c20a2 | [
255,
216,
255,
224,
0,
16,
74,
70,
73,
70,
0,
1,
1,
0,
0,
1,
0,
1,
0,
0,
255,
219,
0,
67,
0,
8,
6,
6,
7,
6,
5,
8,
7,
7,
7,
9,
9,
8,
10,
12,
20,
13,
12,
11,
11,
12,
25,
18,
19,
15,
20,
29,
26,
31,
30,
29,
... | 规模突破千亿元 | text_simplified_chinese | 24 |
c9a4b42520c6b6d20611cdf361c2cf06 | [
255,
216,
255,
224,
0,
16,
74,
70,
73,
70,
0,
1,
1,
0,
0,
1,
0,
1,
0,
0,
255,
219,
0,
67,
0,
8,
6,
6,
7,
6,
5,
8,
7,
7,
7,
9,
9,
8,
10,
12,
20,
13,
12,
11,
11,
12,
25,
18,
19,
15,
20,
29,
26,
31,
30,
29,
... | 字词轻松过关 | text_simplified_chinese | 25 |
4b23cb1501e2a70ef7cde73d660732ed | [
255,
216,
255,
224,
0,
16,
74,
70,
73,
70,
0,
1,
1,
0,
0,
1,
0,
1,
0,
0,
255,
219,
0,
67,
0,
8,
6,
6,
7,
6,
5,
8,
7,
7,
7,
9,
9,
8,
10,
12,
20,
13,
12,
11,
11,
12,
25,
18,
19,
15,
20,
29,
26,
31,
30,
29,
... | 米育种首席专家、省管优秀专家、国务院 | text_simplified_chinese | 26 |
b1dbad2407f4c0e5ad8d6b8a416cfeac | [
255,
216,
255,
224,
0,
16,
74,
70,
73,
70,
0,
1,
1,
0,
0,
1,
0,
1,
0,
0,
255,
219,
0,
67,
0,
8,
6,
6,
7,
6,
5,
8,
7,
7,
7,
9,
9,
8,
10,
12,
20,
13,
12,
11,
11,
12,
25,
18,
19,
15,
20,
29,
26,
31,
30,
29,
... | ·投资建议 | text_simplified_chinese | 27 |
fc8d28861fcedb67eec445a059bc3f78 | [
255,
216,
255,
224,
0,
16,
74,
70,
73,
70,
0,
1,
1,
0,
0,
1,
0,
1,
0,
0,
255,
219,
0,
67,
0,
8,
6,
6,
7,
6,
5,
8,
7,
7,
7,
9,
9,
8,
10,
12,
20,
13,
12,
11,
11,
12,
25,
18,
19,
15,
20,
29,
26,
31,
30,
29,
... | 本报记者 刘伟建 李杰 | text_simplified_chinese | 28 |
794964743a826be117f3ab41725312d6 | [
255,
216,
255,
224,
0,
16,
74,
70,
73,
70,
0,
1,
1,
0,
0,
1,
0,
1,
0,
0,
255,
219,
0,
67,
0,
8,
6,
6,
7,
6,
5,
8,
7,
7,
7,
9,
9,
8,
10,
12,
20,
13,
12,
11,
11,
12,
25,
18,
19,
15,
20,
29,
26,
31,
30,
29,
... | 500 万吨/年；6#20 万立方米储罐项目整体进度大致完成 | text_simplified_chinese | 29 |
6b160461095be297c880e00f9f480a04 | [
255,
216,
255,
224,
0,
16,
74,
70,
73,
70,
0,
1,
1,
0,
0,
1,
0,
1,
0,
0,
255,
219,
0,
67,
0,
8,
6,
6,
7,
6,
5,
8,
7,
7,
7,
9,
9,
8,
10,
12,
20,
13,
12,
11,
11,
12,
25,
18,
19,
15,
20,
29,
26,
31,
30,
29,
... | 用无公害有机肥料，统一栽培技术 | text_simplified_chinese | 30 |
e04d26efbf8828ebb823f06199a1c466 | [
255,
216,
255,
224,
0,
16,
74,
70,
73,
70,
0,
1,
1,
0,
0,
1,
0,
1,
0,
0,
255,
219,
0,
67,
0,
8,
6,
6,
7,
6,
5,
8,
7,
7,
7,
9,
9,
8,
10,
12,
20,
13,
12,
11,
11,
12,
25,
18,
19,
15,
20,
29,
26,
31,
30,
29,
... | 框架已有效整合，但需持续监 | text_simplified_chinese | 31 |
End of preview. Expand in Data Studio
ZH_EN_RecBench
zh_en_rec_bench is a benchmark dataset designed to evaluate the robustness and generalization capabilities of text recognition models across multiple scenarios and both Chinese and English scripts. It is constructed by sampling and manually correcting subsets of data from OmniDocBench and TC-STR, with erroneous ground truth labels revised to ensure high-quality evaluation.
Dataset Overview
This benchmark includes four distinct text recognition scenarios:
| Scene | Number of Samples |
|---|---|
| text_simplified_chinese | 995 |
| text_english | 996 |
| traditional_chinese | 1000 |
| text_en_ch_mixed | 959 |
Data Sources
The following four scenarios are derived from OmniDocBench :
- text_en_ch_mixed
- text_english
- text_simplified_chinese
The following one scenarios are derived from TC-STR :
- traditional_chinese
Dataset Structure
Each data sample consists of:
- image: the image content
- label: the text content within the image
- scene: one of the five predefined scenes
- md5: the unique MD5 hash used as image filename
Usage
To extract the dataset into folders by scene, with each containing image files and a label .txt file, use the following script:
def extract_hf_dataset(parquet_path: str, output_path: str):
"""
Extracts the HF dataset from a Parquet file.
For each scene, creates a folder of images and a label file in the format: <relative_image_path> <label>
"""
import pandas as pd
from pathlib import Path
from tqdm import tqdm
df = pd.read_parquet(parquet_path)
df['scene'] = df['scene'].astype(str)
for scene in tqdm(df['scene'].unique()):
scene_path = Path(output_path) / scene
scene_path.mkdir(parents=True, exist_ok=True)
for index, row in df[df['scene'] == scene].iterrows():
image_path = scene_path / f'{row["md5"]}.png'
image_path.write_bytes(row['image'])
with open(Path(output_path) / f'{scene}.txt', 'a') as f:
f.write(f'{image_path.relative_to(Path(output_path))} {row["label"]}\n')
License
The dataset follows the licenses of its original sources:
- Downloads last month
- 9