Update app.py

app.py

@@ -308,9 +308,10 @@ def _process(img: Image.Image, prompt: str | BoundingBox | None, bg_prompt: str
         try:
             if torch.cuda.is_available():
                 current_device = torch.cuda.current_device()
-                torch.cuda.reset_peak_memory_stats(current_device)
+                with torch.cuda.device(current_device):
+                    torch.cuda.empty_cache()
         except Exception as e:
-            print(f"CUDA memory stats reset failed: {e}")
+            print(f"CUDA memory management failed: {e}")
 
         with torch.cuda.amp.autocast(enabled=torch.cuda.is_available()):
             mask, bbox, time_log = _gpu_process(img, prompt)
@@ -324,7 +325,9 @@ def _process(img: Image.Image, prompt: str | BoundingBox | None, bg_prompt: str
 
         clear_memory()
 
-        return (img, combined, masked_alpha), gr.DownloadButton(value=temp.name, interactive=True)
+        with tempfile.NamedTemporaryFile(delete=False, suffix=".png") as temp:
+            combined.save(temp.name)
+            return (img, combined, masked_alpha), gr.DownloadButton(value=temp.name, interactive=True)
     except Exception as e:
         clear_memory()
         print(f"Processing error: {str(e)}")
@@ -682,10 +685,10 @@ gr.HTML("""
         </div>
     </div>
 """)
+demo.queue(max_size=10)  # 큐 크기 제한
 demo.launch(
     server_name="0.0.0.0",
     server_port=7860,
-    share=True,  # 공개 링크 활성화
-    max_threads=2,
-    allowed_paths=[examples_path]  # 수정된 예제 경로 사용
+    share=False,
+    max_threads=2  # 스레드 수 제한
 )