changes

app.py

@@ -1,6 +1,7 @@
 import gradio as gr
 import os
 from coqui_tts import run_audio_generation_v1
+from metaVoice import run_audio_generation_v2
 import shutil
 os.environ["COQUI_TOS_AGREED"] = "1"
 
@@ -25,8 +26,10 @@ def process_audio(input_text, speaker_audio, speaker_name, option_selected):
         if option_selected == "Xtts_v2":
             # Generate TTS audio using run_audio_generation_v1
             run_audio_generation_v1(input_text)
+        elif option_selected =="metaVoice":
+            run_audio_generation_v2(input_text)
         else:
-            return f"The option '{option_selected}' is not implemented yet."
+            return f"The option is not implemented yet."
         
         # Save the output audio under the speaker's name
         speaker_output_path = f"audio/{speaker_name}.wav"
@@ -48,7 +51,7 @@ with gr.Blocks() as demo:
         speaker_audio = gr.Audio(label="Speaker Audio (to be cloned)", type='filepath', format='wav')
         speaker_name = gr.Textbox(label="Speaker Name", placeholder="Enter the speaker's name.")
 
-    option_selected = gr.Dropdown(choices=["Xtts_v2", "metaVoice(not working at the moment)", "more"], label="Select an Option")
+    option_selected = gr.Dropdown(choices=["Xtts_v2", "metaVoice", "more"], label="Select an Option")
 
     submit_btn = gr.Button("Submit")
 

coqui_tts.py

@@ -32,7 +32,10 @@ def run_audio_generation_v1(new_text,accent='None'):
     for word in new_text_split:
         if len(word)>=2 and word.isupper():
             new_text = new_text.replace(word, " ".join([*word]))
-    
+            
+    models = TTS().list_models()
+    with open('models.txt', 'w') as f:
+        f.writelines(f"{model}\n" for model in models)
     gpu = True if torch.cuda.is_available() else False
     tts = TTS("tts_models/multilingual/multi-dataset/xtts_v2", gpu=gpu) # gpu should be true when server (cuda)
     # if not gpu: