Update app.py

app.py

@@ -2,14 +2,13 @@ import gradio as gr
 import torch
 import huggingsound
 from diffusers import StableDiffusionPipeline, DPMSolverMultistepScheduler
-
+from transformers import SpeechRecognitionModel
 
 def modelo1(audio):
     model = SpeechRecognitionModel("jonatasgrosman/wav2vec2-large-xlsr-53-english")
     transcriptions = model.transcribe(audio)
     return transcriptions
 
-
 def modelo2(text):
     model_id = "stabilityai/stable-diffusion-2-1"
 
@@ -19,12 +18,13 @@ def modelo2(text):
     pipe = pipe.to("cuda")
 
     image = pipe(text).images[0]
-    
+    return image
 
 def execution(audio):
     modelo1res = modelo1(audio)
     modelo2res = modelo2(modelo1res)
-    return modelo1res
+    return modelo2res
 
 if __name__ == "__main__":
+    demo = gr.Interface(fn=execution, inputs="audio", outputs="image")
     demo.launch()