record realtime

app.py

@@ -162,7 +162,7 @@ def transcribe_vi(audio, state_vi="", state_en=""):
     state_vi += beam_search_output + " "
     en_text = translate_vi2en(beam_search_output)
     state_en += en_text + " "
-    return state_vi, state_en, state_vi, state_en
+    return state_vi, state_en
 
 def transcribe_en(audio, state_en="", state_vi=""):
     speech = load_data(audio)
@@ -179,7 +179,7 @@ def transcribe_en(audio, state_en="", state_vi=""):
     state_en += transcription + " "
     vi_text = translate_en2vi(transcription)
     state_vi += vi_text + " "
-    return state_en, state_vi, state_en, state_vi
+    return state_en, state_vi
 
 """Gradio demo"""
 
@@ -207,6 +207,7 @@ with gr.Blocks() as demo:
             translate_button_vien_1.click(lambda text: translate_vi2en(text), inputs=vietnamese_text, outputs=english_out_1)
             gr.Examples(examples=vi_example_text,
                         inputs=[vietnamese_text])
+        
         with gr.TabItem("Speech2text and Vi-En Translation"):
             with gr.Row():
                 with gr.Column():
@@ -215,10 +216,10 @@ with gr.Blocks() as demo:
                 with gr.Column():
                     speech2text_vi1 = gr.Textbox(label="Vietnamese Text")
                     english_out_2 = gr.Textbox(label="English Text")
-                
             translate_button_vien_2.click(lambda vi_voice: inference_vien(vi_voice), inputs=vi_audio, outputs=[speech2text_vi1, english_out_2])
             gr.Examples(examples=vi_example_voice,
                         inputs=[vi_audio])
+        
         with gr.TabItem("Vi-En Realtime Translation"):
             with gr.Row():
                 with gr.Column():
@@ -227,7 +228,7 @@ with gr.Blocks() as demo:
                 with gr.Column():
                     speech2text_vi2 = gr.Textbox(label="Vietnamese Text")
                     english_out_3 = gr.Textbox(label="English Text")
-            vi_audio.change(transcribe_vi, [vi_audio, "state_vi", "state_en"], [speech2text_vi2, english_out_3, "state_vi", "state_en"])
+            vi_audio.change(transcribe_vi, [vi_audio, speech2text_vi2, english_out_3], [speech2text_vi2, english_out_3])
 
 
     with gr.Tabs():
@@ -262,7 +263,7 @@ with gr.Blocks() as demo:
                 with gr.Column():
                     speech2text_en2 = gr.Textbox(label="English Text")
                     vietnamese_out_3 = gr.Textbox(label="Vietnamese Text")
-            en_audio_2.change(transcribe_en, [en_audio_2, "state_en", "state_vi"], [speech2text_en2, vietnamese_out_3, "state_en", "state_vi"])
+            en_audio_2.change(transcribe_en, [en_audio_2, speech2text_en2, vietnamese_out_3], [speech2text_en2, vietnamese_out_3])
 
 if __name__ == "__main__":
     demo.launch()