Update app.py

app.py

@@ -8,6 +8,12 @@ from voicefixer import VoiceFixer
 voicefixer = VoiceFixer()
 import gradio as gr
 import openai
+import torchaudio
+from speechbrain.pretrained import WaveformEnhancement
+enhance_model = WaveformEnhancement.from_hparams(
+    source="speechbrain/mtl-mimic-voicebank",
+    savedir="pretrained_models/mtl-mimic-voicebank",
+)
 
 mes1 = [
     {"role": "system", "content": "You are a TOEFL examiner. Help me improve my oral Englsih and give me feedback."}
@@ -70,9 +76,14 @@ def transcribe(apikey, upload, audio, choice1):
     voicefixer.restore(input="output.wav", # input wav file path
                     output="audio1.wav", # output wav file path
                     cuda=True, # whether to use gpu acceleration
-                    mode = 1) # You can try out mode 0, 1 to find out the best result
+                    mode = 2) # You can try out mode 0, 1, or 2 to find out the best result
+    
+    enhanced = enhance_model.enhance_file("audio1.wav")
+
+    # Saving enhanced signal on disk
+    torchaudio.save("enhanced.wav", enhanced.unsqueeze(0).cpu(), 16000)
 
-    return [result.text, chat_response, "audio1.wav"]
+    return [result.text, chat_response, "enhanced.wav"]
 
 output_1 = gr.Textbox(label="Speech to Text")
 output_2 = gr.Textbox(label="ChatGPT Output")