Update app.py

app.py

@@ -148,7 +148,7 @@ attn_skip_iter = 0
 
 
 def clear_images():
-    return [None, ]*num_inputs + ["", ]*num_inputs 
+    return [None, ]*num_inputs
 
 @spaces.GPU()
 def det_seg_img(image, label):
@@ -541,10 +541,8 @@ if __name__ == "__main__":
                             vlm_btns.append(vlm_btn)
                                 
                     prompt = gr.Textbox(label="Prompt", placeholder="e.g., ENT1 and ENT2")
-                    steps_slider = gr.Slider(minimum=4, maximum=40, step=1, value=8, label="inference steps")
                     gen_btn = gr.Button("Generate", variant="primary")
-                    clear_btn = gr.Button("Clear")
-    
+                    steps_slider = gr.Slider(minimum=4, maximum=40, step=1, value=8, label="inference steps")
                     with gr.Accordion("Advanced Settings", open=False):
 
                         seed = gr.Number(value=42, label="Seed", info="")
@@ -628,6 +626,8 @@ if __name__ == "__main__":
                             double_attention = gr.Checkbox(value=False, label="Double Attention", visible=False)
                             single_attention = gr.Checkbox(value=True, label="Single Attention", visible=False)            
         
+                        clear_btn = gr.Button("Clear Images")
+    
             
                 with gr.Column():
                     output = gr.Image(label="Result")
@@ -708,7 +708,7 @@ if __name__ == "__main__":
             ], 
             outputs=[output, final_text]
         )
-        clear_btn.click(clear_images, outputs=images + captions)
+        clear_btn.click(clear_images, outputs=images)
 
         for i in range(num_inputs):
             face_btns[i].click(det_seg_img, inputs=[images[i], gr.State("A face")], outputs=[images[i]])