fixes

MOCKS.py

@@ -42,7 +42,7 @@ audio testset for evaluation and benchmarking Open-Vocabulary Keyword Spotting (
 
 _BASE_URL = "https://huggingface.co/datasets/voiceintelligenceresearch/MOCKS/tree/main"
 _DL_URL = "https://huggingface.co/datasets/voiceintelligenceresearch/MOCKS/blob/main/"
-_TR_URL = "https://huggingface.co/datasets/voiceintelligenceresearch/MOCKS/blob/main/"
+_TR_URL = "https://huggingface.co/datasets/voiceintelligenceresearch/MOCKS/raw/main/"
 _DL_URLS = {
     "de.MCV": {
         "offline": _DL_URL + "de/MCV/test/offline/data.tar.gz",
@@ -477,8 +477,9 @@ class Mocks(datasets.GeneratorBasedBuilder):
 
         if s_type == "online":
             trans_col_name = "adversary_keyword_transcription"
-
+        logger.info(transcription)
         with open(transcription, encoding="utf-8") as f:
+            print(f)
             for row in f:
                 r = row.split(r"\t")
                 print(row)