change to cpu

app.py

@@ -61,24 +61,28 @@ os.environ["COQUI_TOS_AGREED"] = "1"
 ##print(m)
 #m = model_name
 
+xtts_model = TTS("tts_models/multilingual/multi-dataset/xtts_v2", gpu=False)
+
+
+
 #xtts_model = TTS(model_name, gpu=False)
 #xtts_model.to("cpu") # no GPU or Amd
 #tts.to("cuda") # cuda only
 
-tts_model_name = "tts_models/multilingual/multi-dataset/xtts_v2"
-ModelManager().download_model(tts_model_name)
-tts_model_path = os.path.join(get_user_data_dir("tts"), tts_model_name.replace("/", "--"))
-config = XttsConfig()
-config.load_json(os.path.join(tts_model_path, "config.json"))
-xtts_model = Xtts.init_from_config(config)
-xtts_model.to("cpu")
-xtts_model.load_checkpoint(
-    config,
-    checkpoint_path=os.path.join(tts_model_path, "model.pth"),
-    vocab_path=os.path.join(tts_model_path, "vocab.json"),
-    eval=True,
-    use_deepspeed=True,
-)
+#tts_model_name = "tts_models/multilingual/multi-dataset/xtts_v2"
+#ModelManager().download_model(tts_model_name)
+#tts_model_path = os.path.join(get_user_data_dir("tts"), tts_model_name.replace("/", "--"))
+#config = XttsConfig()
+#config.load_json(os.path.join(tts_model_path, "config.json"))
+#xtts_model = Xtts.init_from_config(config)
+#xtts_model.to("cpu")
+#xtts_model.load_checkpoint(
+#    config,
+#    checkpoint_path=os.path.join(tts_model_path, "model.pth"),
+#    vocab_path=os.path.join(tts_model_path, "vocab.json"),
+#    eval=True,
+#    use_deepspeed=True,
+#)
 #xtts_model.cuda()
 print("Loaded XTTS model")
 
@@ -154,10 +158,18 @@ with gr.Blocks(title="Voice chat with LLM") as demo:
         def handle_speech_generation(sentence, chatbot_history, chatbot_voice):
             if sentence != "":
                 print("Processing sentence")
-                generated_speech = generate_speech_for_sentence(chatbot_history, chatbot_voice, sentence, xtts_model, None, return_as_byte=True)
-                if generated_speech is not None:
-                    _, audio_dict = generated_speech
-                    yield (sentence, chatbot_history, audio_dict["value"])
+                # generate speech by cloning a voice using default settings
+                xtts_model.tts_to_file(text="It took me quite a long time to develop a voice, and now that I have it I'm not going to be silent.",
+                                file_path="output.wav",
+                                speaker_wav=[f"examples/{chatbot_voice}.wav"],
+                                language="en",
+                                split_sentences=True
+                                )
+                yield (sentence, chatbot_history, wave_header_chunk())
+                #generated_speech = generate_speech_for_sentence(chatbot_history, chatbot_voice, sentence, xtts_model, None, return_as_byte=True)
+                #if generated_speech is not None:
+                #    _, audio_dict = generated_speech
+                #    yield (sentence, chatbot_history, audio_dict["value"])
 
         if initial_greeting:
             # Process only the initial greeting if specified