second

app.py

@@ -12,6 +12,8 @@ else:
 print(device)
 repo_id = "cvssp/audioldm-m-full"
 pipe = AudioLDMPipeline.from_pretrained(repo_id, torch_dtype=torch_dtype)
+pipe = pipe.to(device)
+
 #pipe.unet = torch.compile(pipe.unet)
 
 def generate_sound(text): 
@@ -23,8 +25,8 @@ def generate_sound(text):
                      audio_length_in_s=5,
                      negative_prompt = "low quality, average quality").audios
     rate =160000
-    #return rate, audio
-    return gr.make_waveform((rate, waveforms[0]))
+    return rate, audio
+    #return gr.make_waveform((rate, waveforms[0]))
 
 def translate_text(text):
     translator = Translator()
@@ -41,6 +43,7 @@ with demo:
 
         with gr.Column():
             output = gr.Audio()
+            #output = gr.Video(label="Output")
     button.click(generate_sound,text,output)