Update app.py

app.py

@@ -13,20 +13,21 @@ except Exception as e:
     print(f"Error initializing pipeline: {e}")
     pipeline = None
 
-@spaces.GPU
-def process_audio(audio, num_speakers, min_speakers, max_speakers):
+def save_audio(audio):
     if pipeline is None:
         return "Error: Pipeline not initialized"
 
-    # Read the uploaded audio file
-    with open(audio, "rb") as f:
-        audio_data = f.read()
-
     # Save the uploaded audio file to a temporary location
     with open("temp.wav", "wb") as f:
-        f.write(audio_data)
+        f.write(audio)
+
+    return "temp.wav"
+
+@spaces.GPU
+def diarize_audio(temp_file, num_speakers, min_speakers, max_speakers):
+    if pipeline is None:
+        return "Error: Pipeline not initialized"
 
-    # Use the diarization pipeline to process the audio
     try:
         params = {}
         if num_speakers > 0:
@@ -36,12 +37,12 @@ def process_audio(audio, num_speakers, min_speakers, max_speakers):
         if max_speakers > 0:
             params["max_speakers"] = max_speakers
 
-        diarization = pipeline("temp.wav", **params)
+        diarization = pipeline(temp_file, **params)
     except Exception as e:
         return f"Error processing audio: {e}"
 
     # Remove the temporary file
-    os.remove("temp.wav")
+    os.remove(temp_file)
 
     # Return the diarization output
     return str(diarization)
@@ -54,6 +55,11 @@ with gr.Blocks() as demo:
     process_button = gr.Button("Process")
     diarization_output = gr.Textbox(label="Diarization Output")
 
-    process_button.click(fn=process_audio, inputs=[audio_input, num_speakers_input, min_speakers_input, max_speakers_input], outputs=diarization_output)
+    process_button.click(
+        fn=lambda audio, num_speakers, min_speakers, max_speakers: 
+            diarize_audio(save_audio(audio), num_speakers, min_speakers, max_speakers),
+        inputs=[audio_input, num_speakers_input, min_speakers_input, max_speakers_input],
+        outputs=diarization_output
+    )
 
 demo.launch()