Update app.py

app.py

@@ -6,6 +6,7 @@ subprocess.run(["pip", "install", "torchaudio", "--upgrade"])
 import gradio as gr
 from transformers import Wav2Vec2ForCTC, Wav2Vec2Processor
 import torchaudio
+import torch
 
 # Load model and processor
 processor = Wav2Vec2Processor.from_pretrained("jonatasgrosman/wav2vec2-large-xlsr-53-italian")
@@ -14,12 +15,15 @@ model = Wav2Vec2ForCTC.from_pretrained("jonatasgrosman/wav2vec2-large-xlsr-53-it
 # Function to perform ASR on audio data
 def transcribe_audio(audio_data):
     print("Received audio data:", audio_data)  # Debug print
-    if audio_data is None:
-        return "No audio data received."
+    if audio_data is None or len(audio_data) != 2:
+        return "Invalid audio data format."
 
     try:
+        # Extract sample rate and audio waveform from the tuple
+        sample_rate, waveform = audio_data
+
         # Convert audio data to mono and normalize
-        audio_data = torchaudio.transforms.Resample(audio_data[1], 16000)(audio_data[0])
+        audio_data = torchaudio.transforms.Resample(sample_rate, 16000)(waveform)
         audio_data = torchaudio.functional.gain(audio_data, gain_db=5.0)
 
         # Apply custom preprocessing to the audio data if needed