Update app.py

app.py

@@ -167,6 +167,7 @@ def one_shot(image_in,input_text,gender):
               tts.write_to_fp(f)
               f.seek(0)
               sound = AudioSegment.from_file(f.name, format="mp3")
+              os.system(f"rm -rf /content/audio.wav")
               sound.export("/content/audio.wav", format="wav")
               audio_in="/content/audio.wav"
          return one_shot_talking(image_in,audio_in)
@@ -189,7 +190,9 @@ def one_shot(image_in,input_text,gender):
 
           wav, rate = TTSHubInterface.get_prediction(task, model, generator, sample)
           # soundfile.write("/content/audio_before.wav", wav, rate)
+          os.system(f"rm -rf /content/audio_before.wav")
           soundfile.write("/content/audio_before.wav", wav.cpu().clone().numpy(), rate)
+          os.system(f"rm -rf /content/audio.wav")
           cmd='ffmpeg -i /content/audio_before.wav -filter:a "atempo=0.7" -vn /content/audio.wav'
           os.system(cmd)
           audio_in="/content/audio.wav"
@@ -206,7 +209,7 @@ def run():
           image_in = gr.Image(show_label=True, type="filepath",label="Input Image")
           input_text = gr.Textbox(show_label=True,label="Input Text")
           gender = gr.Radio(["Female","Male"],value="Female",label="Gender")
-          video_out = gr.Textbox(show_label=True,label="Output")
+          video_out = gr.Video(show_label=True,label="Output")
         with gr.Row().style(equal_height=True):
           btn = gr.Button("Generate")   
     gr.Markdown(