Datasets:
download
Browse files
README.md
CHANGED
|
@@ -218,6 +218,9 @@ tags:
|
|
| 218 |
This is a large-scale collection of web-crawled documents in 198 world languages, produced by the [HPLT project](https://hplt-project.org/).
|
| 219 |
The source of the data is [Internet Archive](https://archive.org/) and [Common Crawl](https://commoncrawl.org/). For a detailed description of this and previous releases by HPLT, please refer to [our website](https://hplt-project.org/datasets/v3.0).
|
| 220 |
|
|
|
|
|
|
|
|
|
|
| 221 |
## HPLT release v3.0
|
| 222 |
In July 2025, the European HPLT initiative has completed a new release of its monolingual datasets, offering better data quality, more annotations and metadata, and greatly increased volume.
|
| 223 |
HPLT Monolingual Datasets 3.0 comprise some 50 terabytes of compressed data, covering 198 languages.
|
|
|
|
| 218 |
This is a large-scale collection of web-crawled documents in 198 world languages, produced by the [HPLT project](https://hplt-project.org/).
|
| 219 |
The source of the data is [Internet Archive](https://archive.org/) and [Common Crawl](https://commoncrawl.org/). For a detailed description of this and previous releases by HPLT, please refer to [our website](https://hplt-project.org/datasets/v3.0).
|
| 220 |
|
| 221 |
+
**NB: the HPLT datasets are not hosted on HuggingFace! See download instructions below.**
|
| 222 |
+
|
| 223 |
+
|
| 224 |
## HPLT release v3.0
|
| 225 |
In July 2025, the European HPLT initiative has completed a new release of its monolingual datasets, offering better data quality, more annotations and metadata, and greatly increased volume.
|
| 226 |
HPLT Monolingual Datasets 3.0 comprise some 50 terabytes of compressed data, covering 198 languages.
|