fix metadata loading issue

reborn_uasr-librispeech_no_silence_100h.py

@@ -113,12 +113,12 @@ class RebornLibrispeech(datasets.GeneratorBasedBuilder):
     def _split_generators(self, dl_manager):
         
         metadata = dl_manager.download({
-            "train-clean-100": self.config.data_root_url + "/metadata/train-clean-100.tsv",
-            "dev-clean": self.config.data_root_url + "/metadata/dev-clean.tsv",
-            "dev-clean-small": self.config.data_root_url + "/metadata/dev-clean-small.tsv",
-            "dev-other": self.config.data_root_url + "/metadata/dev-other.tsv",
-            "test-clean": self.config.data_root_url + "/metadata/test-clean.tsv",
-            "test-other": self.config.data_root_url + "/metadata/test-other.tsv",
+            "train-clean-100": "metadata/train-clean-100.tsv",
+            "dev-clean": "metadata/dev-clean.tsv",
+            "dev-clean-small": "metadata/dev-clean-small.tsv",
+            "dev-other": "metadata/dev-other.tsv",
+            "test-clean": "metadata/test-clean.tsv",
+            "test-other": "metadata/test-other.tsv",
         })
 
         all_splits = [