trying things with cpu

app.py

@@ -38,7 +38,7 @@ from llama_cpp import Llama
 #from TTS.utils.manage import ModelManager
 
 # Local imports
-from utils import get_sentence #, generate_speech_for_sentence, wave_header_chunk
+from utils import get_sentence, wave_header_chunk #, generate_speech_for_sentence, 
 
 # Load Whisper ASR model
 print("Loading Whisper ASR")
@@ -136,8 +136,8 @@ with gr.Blocks(title="Voice chat with LLM") as demo:
 
     def generate_speech(chatbot_history, chatbot_voice, initial_greeting=False):
         # Start by yielding an initial empty audio to set up autoplay
-        #yield ("", chatbot_history, wave_header_chunk())
-        yield ("", chatbot_history)
+        yield ("", chatbot_history, wave_header_chunk())
+        #yield ("", chatbot_history)
 
         # Helper function to handle the speech generation and yielding process
     #    def handle_speech_generation(sentence, chatbot_history, chatbot_voice):
@@ -159,14 +159,14 @@ with gr.Blocks(title="Voice chat with LLM") as demo:
     #            yield from handle_speech_generation(sentence, chatbot_history, chatbot_voice)
 
     txt_msg = txt_box.submit(fn=add_text, inputs=[chatbot, txt_box], outputs=[chatbot, txt_box], queue=False
-                             )#.then(fn=generate_speech,  inputs=[chatbot,chatbot_voice], outputs=[sentence, chatbot, audio_playback])
+                             ).then(fn=generate_speech,  inputs=[chatbot,chatbot_voice], outputs=[sentence, chatbot, audio_playback])
 
     txt_msg.then(fn=lambda: gr.update(interactive=True), inputs=None, outputs=[txt_box], queue=False)
 
-    #audio_msg = audio_record.stop_recording(fn=add_audio, inputs=[chatbot, audio_record], outputs=[chatbot, txt_box], queue=False
-    #                                        ).then(fn=generate_speech,  inputs=[chatbot,chatbot_voice], outputs=[sentence, chatbot, audio_playback])
+    audio_msg = audio_record.stop_recording(fn=add_audio, inputs=[chatbot, audio_record], outputs=[chatbot, txt_box], queue=False
+                                            ).then(fn=generate_speech,  inputs=[chatbot,chatbot_voice], outputs=[sentence, chatbot, audio_playback])
 
-    #audio_msg.then(fn=lambda: (gr.update(interactive=True),gr.update(interactive=True,value=None)), inputs=None, outputs=[txt_box, audio_record], queue=False)
+    audio_msg.then(fn=lambda: (gr.update(interactive=True),gr.update(interactive=True,value=None)), inputs=None, outputs=[txt_box, audio_record], queue=False)
 
     FOOTNOTE = """
             This Space demonstrates how to speak to an llm chatbot, based solely on open accessible models.
@@ -179,5 +179,5 @@ with gr.Blocks(title="Voice chat with LLM") as demo:
             - Responses generated by chat model should not be assumed correct or taken serious, as this is a demonstration example only
             - iOS (Iphone/Ipad) devices may not experience voice due to autoplay being disabled on these devices by Vendor"""
     gr.Markdown(FOOTNOTE)
-    demo.load(fn=generate_speech, inputs=[chatbot,chatbot_voice, gr.State(value=True)], outputs=[sentence, chatbot])  #outputs=[sentence, chatbot, audio_playback])
+    demo.load(fn=generate_speech, inputs=[chatbot,chatbot_voice, gr.State(value=True)], outputs=[sentence, chatbot, audio_playback])
 demo.queue().launch(debug=True,share=True)