Update app.py

app.py

@@ -34,16 +34,16 @@ def predict(input, history=None):
         history = []
     response, history = model.chat(tokenizer, input, history)
 
-    return history, history, inference(response)
+    return history, history, response
+
+def chinese(text_cn, upload1, VoiceMicrophone1):
 
-def chinese(audio, upload1, VoiceMicrophone1):
-    
     if upload1 is not None:
         
-        tts.voice_conversion_to_file(source_wav=audio, target_wav=upload1, file_path="output0.wav")
+        tts.voice_conversion_to_file(source_wav=inference(text_cn), target_wav=upload1, file_path="output0.wav")
            
     else:
-        tts.voice_conversion_to_file(source_wav=audio, target_wav=VoiceMicrophone1, file_path="output0.wav")
+        tts.voice_conversion_to_file(source_wav=inference(text_cn), target_wav=VoiceMicrophone1, file_path="output0.wav")
 
 
     noisy = enhance_model.load_audio(
@@ -81,17 +81,17 @@ with gr.Blocks() as demo:
     )
     state = gr.State([])
     chatbot = gr.Chatbot([], elem_id="chatbot").style(height=300)
-    audio_res = gr.Audio(label = "自动生成的声音")
+    res = gr.Textbox(lines=1, placeholder="最新的回答在这里", show_label = False).style(container=False)
     with gr.Row():
 #        with gr.Column(scale=4):
         txt = gr.Textbox(label = "说点什么吧(中英皆可)", lines=1)
 #        with gr.Column(scale=1):
         button = gr.Button("开始对话吧")
-    txt.submit(predict, [txt, state], [chatbot, state, audio_res])
-    button.click(predict, [txt, state], [chatbot, state, audio_res])
+    txt.submit(predict, [txt, state], [chatbot, state, res])
+    button.click(predict, [txt, state], [chatbot, state, res])
     
     with gr.Row().style(mobile_collapse=False, equal_height=True):
-        inp3 = audio_res
+        inp3 = res
         inp4 = gr.Audio(source="upload", label = "请上传您喜欢的声音(wav/mp3文件)；长语音(90s左右)效果更好", type="filepath")
         inp5 = gr.Audio(source="microphone", type="filepath", label = '请用麦克风上传您喜欢的声音，与文件上传二选一即可')
         btn1 = gr.Button("用喜欢的声音听一听吧(中文)")