Update audioset.py

audioset.py

@@ -167,6 +167,7 @@ class AudioSet(datasets.GeneratorBasedBuilder):
             try:
                 metadata = torchaudio.info(wav_path)
                 duration = math.ceil(metadata.num_frames / metadata.sample_rate) # in seconds
+                print(duration)
                 if duration > 0:
                     yield guid, {
                         "id": str(guid),