Update app.py

app.py

@@ -11,8 +11,14 @@ pipe1 = pipeline("automatic-speech-recognition", model="openai/whisper-base")
 #pipe3.to("cuda" if torch.cuda.is_available() else "cpu")
 
 def audio_to_image(audio):
+        # Load the audio file and convert it to a numpy array
+        audio_data, _ = sf.read(audio)  # Load audio file
+        audio_array = np.array(audio_data)  # Convert to numpy array
     
-        transcription = pipe1(audio)
+        # Transcribe the audio input
+        transcription = pipe1(audio_array)
+    
+        #transcription = pipe1(audio)
 
         transcription_text = transcription['text']