Update main.py

main.py

@@ -53,6 +53,27 @@ def generate_emotional_speech(text, emotion):
     "shame": {"pitch": 0.8, "speed": 0.85},      # Quiet, subdued tone
 
 }
+    
+import librosa
+import soundfile as sf
+
+def adjust_pitch(audio_path, pitch_factor):
+    # Load audio
+    y, sr = librosa.load(audio_path)
+    # Adjust pitch
+    y_shifted = librosa.effects.pitch_shift(y, sr, n_steps=pitch_factor)
+    # Save adjusted audio
+    sf.write(audio_path, y_shifted, sr)
+
+def adjust_speed(audio_path, speed_factor):
+    # Load the audio file
+    y, sr = librosa.load(audio_path)
+
+    # Adjust the speed (this alters the duration of the audio)
+    y_speeded = librosa.effects.time_stretch(y, speed_factor)
+
+    # Save the adjusted audio
+    sf.write(audio_path, y_speeded, sr)
 
     # Retrieve pitch and speed based on detected emotion
     settings = emotion_settings.get(emotion, {"pitch": 1.0, "speed": 1.0})
@@ -250,27 +271,6 @@ save_path = "/content/drive/My Drive/fine_tuned_tacotron2.pth"
 # Save the model's state dictionary using torch.save
 torch.save(model.state_dict(), save_path)
 
-import librosa
-import soundfile as sf
-
-def adjust_pitch(audio_path, pitch_factor):
-    # Load audio
-    y, sr = librosa.load(audio_path)
-    # Adjust pitch
-    y_shifted = librosa.effects.pitch_shift(y, sr, n_steps=pitch_factor)
-    # Save adjusted audio
-    sf.write(audio_path, y_shifted, sr)
-
-def adjust_speed(audio_path, speed_factor):
-    # Load the audio file
-    y, sr = librosa.load(audio_path)
-
-    # Adjust the speed (this alters the duration of the audio)
-    y_speeded = librosa.effects.time_stretch(y, speed_factor)
-
-    # Save the adjusted audio
-    sf.write(audio_path, y_speeded, sr)
-
 
 """Set up the Gradio interface"""