Update README.md
Browse files
README.md
CHANGED
|
@@ -4,7 +4,19 @@ This enables you to load any raw mono corpus from [opus.nlpl.eu](https://opus.nl
|
|
| 4 |
|
| 5 |
To use it, you need the name of the corpus, the version, and the target language code. The corpus name and version are provided in one string seperated by space (e.g. 'News-Commentary v16'). All of these can be found on [opus.nlpl.eu](https://opus.nlpl.eu/).
|
| 6 |
|
|
|
|
|
|
|
| 7 |
## Example:
|
| 8 |
```python
|
| 9 |
dataset = load_dataset('badranx/opus_raw', corpus="News-Commentary v16", lang="de")
|
| 10 |
```
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 4 |
|
| 5 |
To use it, you need the name of the corpus, the version, and the target language code. The corpus name and version are provided in one string seperated by space (e.g. 'News-Commentary v16'). All of these can be found on [opus.nlpl.eu](https://opus.nlpl.eu/).
|
| 6 |
|
| 7 |
+
I didn't provide any default dataset. You must provide two parameters as configurations: corpus and lang, see example below.
|
| 8 |
+
|
| 9 |
## Example:
|
| 10 |
```python
|
| 11 |
dataset = load_dataset('badranx/opus_raw', corpus="News-Commentary v16", lang="de")
|
| 12 |
```
|
| 13 |
+
## Structure
|
| 14 |
+
The structure is simple.
|
| 15 |
+
```python
|
| 16 |
+
{
|
| 17 |
+
"id": datasets.Value("string"),
|
| 18 |
+
"text": datasets.Value("string"),
|
| 19 |
+
}
|
| 20 |
+
```
|
| 21 |
+
|
| 22 |
+
"text" can be one or more sentences, but not more than a paragraph.
|