dataloader

nbnn_language_detection.py

@@ -1,4 +1,4 @@
-from datasets import DatasetBuilder, DatasetInfo, SplitGenerator, load_dataset
+from datasets import DatasetBuilder, DatasetInfo, SplitGenerator
 from datasets.features import Features, Value
 import json
 
@@ -16,15 +16,17 @@ class MyDataset(DatasetBuilder):
         )
 
     def _split_generators(self, dl_manager):
-        files = {
-            'train': 'train.jsonl',
-            'dev': 'dev.jsonl',
-            'test': 'test.jsonl'
+        urls = {
+            'train': 'https://huggingface.co/datasets/NbAiLab/nbnn_language_detection/resolve/main/train.jsonl',
+            'dev': 'https://huggingface.co/datasets/NbAiLab/nbnn_language_detection/resolve/main/dev.jsonl',
+            'test': 'https://huggingface.co/datasets/NbAiLab/nbnn_language_detection/resolve/main/test.jsonl',
         }
+        
+        downloaded_files = dl_manager.download(urls)
 
         return [
-            SplitGenerator(name=split, gen_kwargs={'filepath': path})
-            for split, path in files.items()
+            SplitGenerator(name=split, gen_kwargs={'filepath': downloaded_files[split]})
+            for split in urls.keys()
         ]
 
     def _generate_examples(self, filepath):