Update app.py

app.py

@@ -79,6 +79,9 @@ def run_asr_ctc(input_audio: str, target_language: str) -> str:
     # Ensure shape [B x T]
     if len(audio_tensor.shape) == 1:
         audio_tensor = audio_tensor.unsqueeze(0)  # Add batch dimension if missing
+
+    if audio_tensor.ndim > 1:
+        audio_tensor = audio_tensor.squeeze(0)
     
     # Resample to 16kHz
     audio_tensor = torchaudio.functional.resample(audio_tensor, orig_freq=orig_freq, new_freq=16000)
@@ -114,6 +117,9 @@ def run_asr_rnnt(input_audio: str, target_language: str) -> str:
     # Ensure shape [B x T]
     if len(audio_tensor.shape) == 1:
         audio_tensor = audio_tensor.unsqueeze(0)  # Add batch dimension if missing
+
+    if audio_tensor.ndim > 1:
+        audio_tensor = audio_tensor.squeeze(0)
     
     # Resample to 16kHz
     audio_tensor = torchaudio.functional.resample(audio_tensor, orig_freq=orig_freq, new_freq=16000)