Update app.py

app.py

@@ -40,13 +40,14 @@ def handle_audio(audio):
     audio = audio[1]
     input_values = wav2vec2_processor(audio, sampling_rate=16_000, return_tensors="pt").input_values
     # Convert to the expected tensor type
-    input_values = input_values.to(torch.float64)
+    input_values = input_values.to(torch.float32)
     logits = wav2vec2_model(input_values).logits
     predicted_ids = torch.argmax(logits, dim=-1)
     transcriptions = wav2vec2_processor.decode(predicted_ids[0])
     return handle_text(transcriptions)
 
 
+
 def chatbot(text, img, audio):
     text_output = handle_text(text) if text is not None else ''
     img_output = handle_image(img) if img is not None else ''