Datasets:
BramVanroy
commited on
Update README.md
Browse files
README.md
CHANGED
@@ -26,4 +26,32 @@ configs:
|
|
26 |
data_files:
|
27 |
- split: train
|
28 |
path: data/train-*
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
29 |
---
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
26 |
data_files:
|
27 |
- split: train
|
28 |
path: data/train-*
|
29 |
+
license: cc0-1.0
|
30 |
+
task_categories:
|
31 |
+
- text-generation
|
32 |
+
language:
|
33 |
+
- nl
|
34 |
+
pretty_name: HPLT Dutch cleaned v1.2
|
35 |
+
size_categories:
|
36 |
+
- 100B<n<1T
|
37 |
---
|
38 |
+
|
39 |
+
# HPLT Dutch cleaned v1.2
|
40 |
+
|
41 |
+
- Data creator: [High Performance Language Technologies](https://hplt-project.org/)
|
42 |
+
- Data URL: https://hplt-project.org/datasets/v1.2
|
43 |
+
- Technical data description: https://hplt-project.org/HPLT_D2_1___Initial_release_of_monolingual_and_parallel_data_sets-1.pdf
|
44 |
+
|
45 |
+
## Fields
|
46 |
+
|
47 |
+
- `id`: Document ID
|
48 |
+
- `document_lang`: Document language identified by CLD2 during the WARC extraction process.
|
49 |
+
- `scores`: Language identification scores for each paragraph in the document.
|
50 |
+
- `langs`: Language with highest score for each paragraph in the document.
|
51 |
+
- `text`: The document's text (a concatenation of newline-separated paragraphs).
|
52 |
+
- `url`: Document URL.
|
53 |
+
- `collection`: Collection name.
|
54 |
+
|
55 |
+
## Data removal
|
56 |
+
|
57 |
+
Found data that you would like removed in the next release? Contact [the data creators](mailto:[email protected]).
|