Dataset Viewer
Auto-converted to Parquet Duplicate
text
stringlengths
300
301k
title
stringlengths
1
208
year
int16
2.01k
2.01k
snapshot
stringclasses
1 value
length_chars
int32
300
301k
raw_length_chars
int32
500
320k
id
stringlengths
12
12
processed_at
stringlengths
20
27
''Anarchism'' (from Greek ἀν (without) + ἄρχειν (to rule) + ισμός (from stem -ιζειν), "without archons," "without rulers") is a political philosophy encompassing theories and attitudes which reject compulsory government (the state) and support its elimination, often due to a wider rejection of involuntary or permanent ...
Anarchism
2,008
20080103
34,959
66,363
b0d3db46319c
2026-05-31T08:27:28.302899Z
* * Connections to autism resources at the National Dissemination Center for Children with Disabilities Category:Autism Category:Pervasive Developmental Disorders Category:Childhood psychiatric disorders Category:Communication disorders Category:Neurological disorders ar:توحّد bn:আত্মসংবৃতি bs:Autizam bg:Аутизъм ca:...
Autism
2,008
20080103
848
87,265
ca6c952741fc
2026-05-31T08:27:28.306794Z
The ''albedo'' of an object is the extent to which it diffusely reflects light, defined as the ratio of diffusely reflected to incident electromagnetic radiation. It is a unitless measure indicative of a surface's or body's diffuse reflectivity. The word is derived from Latin 'albedo' "whiteness", in turn from 'albus' ...
Albedo
2,008
20080103
14,213
18,397
c5801a81050d
2026-05-31T08:27:28.307787Z
''Abu Dhabi'' ( '', literally "Father of Gazelle") is the capital and second largest city of the United Arab Emirates. It is also the capital and largest city of the emirate of Abu Dhabi, which is the largest of the seven emirates of the United Arab Emirates by size. It was said by CNN to be the richest city in the wo...
Abu Dhabi
2,008
20080103
13,791
18,103
da4cfad14f0a
2026-05-31T08:27:28.308811Z
The letter ''A'' is the first letter in the Latin alphabet. Its name in English is 'a' (), plural 'aes,' 'a's, or 'a's'. == History == The letter A can be traced to a pictogram of an ox head in Egyptian hieroglyph or the Proto-semitic alphabet. By about 1600 B.C., the Phoenician alphabet's letter had a linear form t...
A
2,008
20080103
4,244
7,014
d7c093115028
2026-05-31T08:27:28.310480Z
-wide meteorite impact crater is located in Elmore County, just north of Montgomery. This is the Wetumpka crater, which is the site of "Alabama's greatest natural disaster". A -wide meteorite hit the area about 80 million years ago. The hills just east of downtown Wetumpka showcase the eroded remains of the impact cra...
Alabama
2,008
20080103
27,801
73,231
970fa7ebdd9a
2026-05-31T08:27:28.315085Z
thumb|right|250px|'The Wrath of Achilles', by François-Léon Benouville (1821–1859) (Musée Fabre) In Greek mythology, ''Achilles'' (also ''Akhilleus'' or ''Achilleus''; Ancient Greek: ) was a Greek hero of the Trojan War, the central character and greatest warrior of Homer's 'Iliad', which takes for its theme the ...
Achilles
2,008
20080103
30,311
36,430
230fe821cca9
2026-05-31T08:27:28.318545Z
| birth_place =Hardin County, Kentucky | death_date = | death_place =Washington, D.C. | spouse = Mary Todd Lincoln | children = Edward Lincoln, Robert Todd Lincoln, Willie Lincoln, Tad Lincoln | occupation=Lawyer | religion = attended churches, but never officially acquired membership in a church | party = Whig, Repub...
Abraham Lincoln
2,008
20080103
63,727
95,202
f9c9b812c3ce
2026-05-31T08:27:28.323702Z
"'Aristotélēs'\n| image_name = Aristoteles Louvre.jpg\n| color = #B0C4DE\n| region = Western philo(...TRUNCATED)
Aristotle
2,008
20080103
48,238
60,626
b31297d5d4c7
2026-05-31T08:27:28.327762Z
": 'This article is about the Gershwin composition. For the 1951 musical starring Gene Kelly, see An(...TRUNCATED)
An American in Paris
2,008
20080103
3,783
4,237
6f8f8163d81c
2026-05-31T08:27:28.328281Z
End of preview. Expand in Data Studio

Wikipedia 2008 & 2010 Dataset

A curated collection of 4.79 million Wikipedia articles spanning the 2008 and 2010 snapshot releases, cleaned and compressed for efficient large-scale language model pretraining. This dataset preserves the raw encyclopedic knowledge of two distinct eras of Wikipedia, making it valuable for temporal analysis, knowledge evolution research, and foundation model training.

Dataset Statistics

Split Articles Compressed Size Raw JSONL Size
wiki_2008 1,903,473 2.43 GB 5.70 GB
wiki_2010 2,885,860 3.63 GB 8.65 GB
Total 4,789,333 6.05 GB 14.35 GB

Format & Schema

Both splits are stored as Parquet files with zstd compression, offering ~60% space savings over raw JSONL while enabling fast columnar reads and streaming. The schema includes:

  • text — cleaned article body content
  • title — article title
  • year — snapshot year (2008 or 2010)
  • snapshot — specific dump identifier
  • length_chars — cleaned character count
  • raw_length_chars — original character count before stripping
  • id — unique hexadecimal article identifier
  • processed_at — ISO timestamp of conversion

Usage

Load and stream directly from Hugging Face without downloading the full 6GB:

from datasets import load_dataset

# Stream 2008 articles
ds_2008 = load_dataset("adhyanshaa/wikipideia-2008-2010", split="wiki_2008", streaming=True)

# Or load 2010 split
ds_2010 = load_dataset("adhyanshaa/wikipideia-2008-2010", split="wiki_2010", streaming=True)

for article in ds_2008:
    print(article["title"], len(article["text"]))
    break

Source & Processing

Original data sourced from Wikipedia XML dumps (2008 and 2010 snapshots). Articles were extracted, wiki markup stripped, and metadata preserved. The conversion pipeline uses streaming PyArrow with atomic writes and row-count validation to ensure zero data loss during the JSONL-to-Parquet transformation.

License

This dataset is released under CC BY-SA 4.0 to remain compliant with the original Wikipedia content license. When using this data for model training or redistribution, attribution to Wikipedia and the Wikimedia Foundation is required.


Wikipedia 2008 & 2010 Dataset

A curated collection of 4.79 million Wikipedia articles spanning the 2008 and 2010 snapshot releases, cleaned and compressed for efficient large-scale language model pretraining. This dataset preserves the raw encyclopedic knowledge of two distinct eras of Wikipedia, making it valuable for temporal analysis, knowledge evolution research, and foundation model training.

Dataset Statistics

Split Articles Compressed Size Raw JSONL Size
wiki_2008 1,903,473 2.43 GB 5.70 GB
wiki_2010 2,885,860 3.63 GB 8.65 GB
Total 4,789,333 6.05 GB 14.35 GB

Format & Schema

Both splits are stored as Parquet files with zstd compression, offering ~60% space savings over raw JSONL while enabling fast columnar reads and streaming. The schema includes:

  • text — cleaned article body content
  • title — article title
  • year — snapshot year (2008 or 2010)
  • snapshot — specific dump identifier
  • length_chars — cleaned character count
  • raw_length_chars — original character count before stripping
  • id — unique hexadecimal article identifier
  • processed_at — ISO timestamp of conversion

Usage

Load and stream directly from Hugging Face without downloading the full 6GB:

from datasets import load_dataset

# Stream 2008 articles
ds_2008 = load_dataset("adhyanshaa/wikipideia-2008-2010", split="wiki_2008", streaming=True)

# Or load 2010 split
ds_2010 = load_dataset("adhyanshaa/wikipideia-2008-2010", split="wiki_2010", streaming=True)

for article in ds_2008:
    print(article["title"], len(article["text"]))
    break

Source & Processing

Original data sourced from Wikipedia XML dumps (2008 and 2010 snapshots). Articles were extracted, wiki markup stripped, and metadata preserved. The conversion pipeline uses streaming PyArrow with atomic writes and row-count validation to ensure zero data loss during the JSONL-to-Parquet transformation.

License

This dataset is released under CC BY-SA 4.0 to remain compliant with the original Wikipedia content license. When using this data for model training or redistribution, attribution to Wikipedia and the Wikimedia Foundation is required.

Downloads last month
77