Update app.py

app.py

@@ -1,5 +1,5 @@
 import gradio as gr
-
+from moviepy.editor import *
 from share_btn import community_icon_html, loading_icon_html, share_js
 
 caption = gr.Blocks.load(name="spaces/SRDdev/Image-Caption")
@@ -8,7 +8,10 @@ audio_gen = gr.Blocks.load(name="spaces/haoheliu/audioldm-text-to-audio-generati
 def infer(image_input):
     cap = caption(image_input, fn_index=0)
     sound = audio_gen(cap, 5, 2.5, 45, 3, fn_index=0)
-    return gr.Textbox.update(value=cap, visible=True), sound, gr.Group.update(visible=True)
+    FILETOCONVERT = AudioFileClip(sound)
+    FILETOCONVERT.write_audiofile("audio.wav")
+    FILETOCONVERT.close()
+    return gr.Textbox.update(value=cap, visible=True), "audio.wav", gr.Group.update(visible=True)
 
 title = """
     <div style="text-align: center; max-width: 700px; margin: 0 auto;">
@@ -47,7 +50,7 @@ with gr.Blocks(css="style.css") as demo:
     
         input_img = gr.Image(type="filepath", elem_id="input-img")
         caption_output = gr.Textbox(label="Caption", lines=1, visible=False, elem_id="text-caption")
-        sound_output = gr.Video(label="Result", elem_id="sound-output")
+        sound_output = gr.Audio(label="Result", elem_id="sound-output")
         
         generate = gr.Button("Generate SFX from Image")