Datasets:
updated readme
Browse files
README.md
CHANGED
@@ -264,6 +264,36 @@ We provide the un-annotated additional data stats below:
|
|
264 |
| English-New York | 6,454 |
|
265 |
| **Total** | **55,702** |
|
266 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
267 |
### License
|
268 |
The dataset is distributed under the Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International License (CC BY-NC-SA 4.0). The full license text can be found in the accompanying licenses_by-nc-sa_4.0_legalcode.txt file.
|
269 |
|
|
|
264 |
| English-New York | 6,454 |
|
265 |
| **Total** | **55,702** |
|
266 |
|
267 |
+
### How to download data
|
268 |
+
|
269 |
+
```
|
270 |
+
import os
|
271 |
+
import json
|
272 |
+
from datasets import load_dataset
|
273 |
+
|
274 |
+
|
275 |
+
dataset_names = ['arabic_qa', 'assamese_in', 'bangla_bd', 'bangla_in', 'english_bd',
|
276 |
+
'english_qa', 'hindi_in', 'nepali_np', 'turkish_tr']
|
277 |
+
base_dir="./MNQA/"
|
278 |
+
|
279 |
+
for dname in dataset_names:
|
280 |
+
output_dir = os.path.join(base_dir, dname)
|
281 |
+
# load each language
|
282 |
+
dataset = load_dataset("QCRI/MultiNativQA", name=dname)
|
283 |
+
# Save the dataset to the specified directory. This will save all splits to the output directory.
|
284 |
+
dataset.save_to_disk(output_dir)
|
285 |
+
# iterate over splits to save the data into json format
|
286 |
+
for split in ['train','dev','test']:
|
287 |
+
data = []
|
288 |
+
if split not in dataset:
|
289 |
+
continue
|
290 |
+
for idx, item in enumerate(dataset[split]):
|
291 |
+
data.append(item)
|
292 |
+
output_file = os.path.join(output_dir, f"{split}.json")
|
293 |
+
with open(output_file, 'w', encoding='utf-8') as f:
|
294 |
+
json.dump(data, f, ensure_ascii=False, indent=4)
|
295 |
+
```
|
296 |
+
|
297 |
### License
|
298 |
The dataset is distributed under the Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International License (CC BY-NC-SA 4.0). The full license text can be found in the accompanying licenses_by-nc-sa_4.0_legalcode.txt file.
|
299 |
|