diff --git a/OpenOrca.py b/OpenOrca.py
index 3d4f215..ecdfcd6 100644
--- a/OpenOrca.py
+++ b/OpenOrca.py
@@ -15,14 +15,14 @@ class CustomDataset(DatasetBuilder):
         )
 
     def _split_generators(self, dl_manager):
-        base_path = 'path_to_your_data'
+        base_path = dl_manager.download_and_extract('https://huggingface.co/datasets/Open-Orca/OpenOrca')
         folders = ['001-cot', '002-flan', '003-flan-1m', '004-flan1m-aug-shuf', '005-flan-5m', 
                    '006-flan-chatgpt', '007-gpt4_100k', '008-niv', '009-t0']  # add more as needed
         split_generators = []
         for folder in folders:
             split_generators.extend([
-               SplitGenerator(name=f'{folder.replace("-", "_")}_train', gen_kwargs={"filepath": f'{folder}/train.jsonl'}),
-               SplitGenerator(name=f'{folder.replace("-", "_")}_test', gen_kwargs={"filepath": f'{folder}/test.jsonl'}),
+               SplitGenerator(name=f'{folder.replace("-", "_")}_train', gen_kwargs={"filepath": os.path.join(base_path, f'{folder}/train.jsonl')}),
+               SplitGenerator(name=f'{folder.replace("-", "_")}_test', gen_kwargs={"filepath": os.path.join(base_path, f'{folder}/test.jsonl')}),
             ])
         return split_generators
 
@@ -36,3 +36,4 @@ class CustomDataset(DatasetBuilder):
                     'question': data['question'],
                     'response': data['response']
                 }
+