Workaround for duplicated ID's issue

dolma.py

@@ -18,6 +18,7 @@
 import gzip
 import json
 import os
+import uuid
 from typing import List
 
 import datasets
@@ -131,7 +132,7 @@ class Dolma(datasets.GeneratorBasedBuilder):
             with gzip.open(fn, mode="rt", encoding="utf-8") as f:
                 for line in f:
                     row = json.loads(line)
-                    yield row["id"], {
+                    yield str(uuid.uuid4()), {
                         "id": row["id"],
                         "text": row["text"],
                         "added": row.get("added", ""),