Datasets:

Modalities:
Text
Formats:
parquet
Languages:
Thai
Libraries:
Datasets
Dask
License:
wannaphong's picture
Update README.md
8af1554 verified
---
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
dataset_info:
features:
- name: title
dtype: string
- name: text
dtype: string
- name: source
dtype: string
- name: __index_level_0__
dtype: int64
splits:
- name: train
num_bytes: 1255848857
num_examples: 196533
download_size: 429985288
dataset_size: 1255848857
license: cc-by-sa-3.0
task_categories:
- text-generation
language:
- th
size_categories:
- 100K<n<1M
---
# Dataset Card for "thai-wiki-dataset-v3"
This dataset collects all Thai Wikimedia project that cleaned all text for Thai language. Example: [Wikipedia](https://th.wikipedia.org/), [Wikiquote](https://th.wikiquote.org/), [Wikibooks](https://th.wikibooks.org/), [Wikisource](https://th.wikisource.org), and [Wiktionary](https://th.wiktionary.org/).
**Use cause**: RAG, and pretraining model.
License: cc-by-sa-3.0