Update app.py

app.py

@@ -6,7 +6,7 @@ from share_btn import community_icon_html, loading_icon_html, share_js
 
 token = os.environ.get('HF_TOKEN')
 caption = gr.Blocks.load(name="spaces/laion/CoCa")
-audio_gen = gr.Blocks.load(name="spaces/haoheliu/audioldm-text-to-audio-generation)
+audio_gen = gr.Blocks.load(name="spaces/haoheliu/audioldm-text-to-audio-generation")
 
 ph_message="If you're not happy with sound result, you can manually describe the scene depicted in your image :)"
 
@@ -19,7 +19,7 @@ def input_changes(input_img):
         print("CoCa caption: '" + cap + "' • ")
         ph_update = "CoCa caption: '" + cap + "' • "
         
-        return manual_cap.update(value="",placeholder=f"{ph_update}{ph_message}"), caption_output.update(value=cap), debug.update(value=None)
+        return manual_cap.update(value="",placeholder=f"{ph_update}{ph_message}"), caption_output.update(value=cap), sound_output.update(value=None)
  
 def infer(image_input, manual_caption, duration_in, seed, caption_output):
     
@@ -36,9 +36,14 @@ def infer(image_input, manual_caption, duration_in, seed, caption_output):
     
     sound = audio_gen(cap, duration_in, 2.5, seed, 3, fn_index=0)
     print(sound)
-    
+  
+
+    video = VideoFileClip(sound)
+    audio = video.audio
+
+    audio.write_audiofile("sound.mp3")
     #return cap, sound[1], gr.Textbox.update(placeholder=f"{ph_update}{ph_message}"), gr.Group.update(visible=True)
-    return cap, sound, gr.Group.update(visible=True)
+    return cap, "sound.mp3", gr.Group.update(visible=True)
 
 title = """
     <div style="text-align: center; max-width: 700px; margin: 0 auto;">
@@ -104,8 +109,8 @@ with gr.Blocks(css="style.css") as demo:
                 seed_in = gr.Slider(label="Seed", value=440, minimum=45, maximum=10000, step=1)
         
         caption_output = gr.Textbox(label="Caption", visible=False, elem_id="text-caption")
-        #sound_output = gr.Audio(label="Result", elem_id="sound-output")
-        debug = gr.Textbox()
+        sound_output = gr.Audio(label="Result", elem_id="sound-output")
+        #debug = gr.Textbox()
         generate = gr.Button("Generate SFX from Image")
 
         with gr.Group(elem_id="share-btn-container", visible=False) as share_group:
@@ -120,7 +125,7 @@ with gr.Blocks(css="style.css") as demo:
     
 
     
-    generate.click(infer, inputs=[input_img, manual_cap, duration_in, seed_in, caption_output], outputs=[caption_output, debug, share_group], api_name="i2fx")
+    generate.click(infer, inputs=[input_img, manual_cap, duration_in, seed_in, caption_output], outputs=[caption_output, sound_output, share_group], api_name="i2fx")
     share_button.click(None, [], [], _js=share_js)
 
 demo.queue(max_size=32).launch(debug=True)