update app files

README.md

@@ -1,10 +1,10 @@
 ---
-title: Voice Chat With Llm
+title: Voice Chat With LLM
 emoji: 👾
-colorFrom: purple
-colorTo: gray
+colorFrom: blue
+colorTo: red
 sdk: gradio
-sdk_version: 4.26.0
+sdk_version: 3.48.0
 app_file: app.py
 pinned: false
 ---

app.py

@@ -103,7 +103,7 @@ with gr.Blocks(title="Voice chat with LLM") as demo:
             container=False,
             interactive=True,
         )
-        audio_record = gr.Audio(sources=["microphone"], type="filepath", scale=4)
+        audio_record = gr.Audio(source="microphone", type="filepath", scale=4)
 
     # Define generated audio playback component 
     with gr.Row():
@@ -121,7 +121,7 @@ with gr.Blocks(title="Voice chat with LLM") as demo:
     def add_text(chatbot_history, text):
         chatbot_history = [] if chatbot_history is None else chatbot_history
         chatbot_history = chatbot_history + [(text, None)]
-        return chatbot_history, gr.Textbox(value="", interactive=False)
+        return chatbot_history, gr.update(value="", interactive=False)
     
     # Will be triggered on voice submit (will transribe and send to generate_speech)
     def add_audio(chatbot_history, audio):
@@ -131,7 +131,7 @@ with gr.Blocks(title="Voice chat with LLM") as demo:
         text = list(response)[0].text.strip()
         print("Transcribed text:", text)
         chatbot_history = chatbot_history + [(text, None)]
-        return chatbot_history, gr.Textbox(value="", interactive=False)
+        return chatbot_history, gr.update(value="", interactive=False)
 
     def generate_speech(chatbot_history, chatbot_voice, initial_greeting=False):
         # Start by yielding an initial empty audio to set up autoplay
@@ -159,12 +159,12 @@ with gr.Blocks(title="Voice chat with LLM") as demo:
     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])
 
-    txt_msg.then(fn=lambda: gr.Textbox(interactive=True), inputs=None, outputs=[txt_box], queue=False)
+    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.then(fn=lambda: (gr.Audio(interactive=True),gr.Audio(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.

utils.py

@@ -387,9 +387,9 @@ def generate_speech_for_sentence(history, chatbot_voice, sentence, xtts_model, x
                         f.setframerate(24000)
                         f.writeframes(wav_bytestream)
                            
-                    return (history , gr.Audio(value=audio_unique_filename, autoplay=True))
+                    return (history , gr.Audio.update(value=audio_unique_filename, autoplay=True))
                 else:
-                    return (history , gr.Audio(value=wav_bytestream, autoplay=True))
+                    return (history , gr.Audio.update(value=wav_bytestream, autoplay=True))
     except RuntimeError as e:
         if "device-side assert" in str(e):
             # cannot do anything on cuda device side error, need tor estart