pr_Mais
commited on
Commit
•
42bba5d
1
Parent(s):
7161f6e
docs: update README
Browse files
README.md
CHANGED
@@ -1,3 +1,44 @@
|
|
1 |
# Arabic Wiki Dataset
|
2 |
|
3 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
# Arabic Wiki Dataset
|
2 |
|
3 |
+
## Dataset Summary
|
4 |
+
This dataset is extracted using [`wikiextractor`](https://github.com/attardi/wikiextractor) tool, from [Wikipedia Arabic pages](https://dumps.wikimedia.org/arwiki/).
|
5 |
+
|
6 |
+
## Supported Tasks and Leaderboards
|
7 |
+
Intended to train **Arabic** language models on MSA (Modern Spoken Arabic).
|
8 |
+
|
9 |
+
## Dataset Structure
|
10 |
+
The dataset is structured into 2 folders:
|
11 |
+
- `arwiki_20211213_txt`: dataset is divided into subfolders each of which contains no more than 100 documents.
|
12 |
+
- `arwiki_20211213_txt_single`: all documents merged together in a single txt file.
|
13 |
+
|
14 |
+
## Dataset Statistics
|
15 |
+
|
16 |
+
#### Extracts from **December 13, 2021**:
|
17 |
+
|
18 |
+
| documents | vocabulary | words |
|
19 |
+
| --- | --- | --- |
|
20 |
+
| 1,136,455 | 5,446,560 | 175,566,016 |
|
21 |
+
|
22 |
+
## Usage
|
23 |
+
Load all dataset from the single txt file:
|
24 |
+
```python
|
25 |
+
load_dataset('CALM/arwiki',
|
26 |
+
data_files='arwiki_2021_txt_single/arwiki_20211213.txt')
|
27 |
+
|
28 |
+
# OR with stream
|
29 |
+
|
30 |
+
load_dataset('CALM/arwiki',
|
31 |
+
data_files='arwiki_2021_txt_single/arwiki_20211213.txt',
|
32 |
+
streaming=True)
|
33 |
+
```
|
34 |
+
Load a smaller subset from the individual txt files:
|
35 |
+
```python
|
36 |
+
load_dataset('CALM/arwiki',
|
37 |
+
data_files='arwiki_2021_txt/AA/arwiki_20211213_1208.txt')
|
38 |
+
|
39 |
+
# OR with stream
|
40 |
+
|
41 |
+
load_dataset('CALM/arwiki',
|
42 |
+
data_files='arwiki_2021_txt/AA/arwiki_20211213_1208.txt',
|
43 |
+
streaming=True)
|
44 |
+
```
|