diff --git a/iopaint/__init__.py b/iopaint/__init__.py
index 2be8d8b..d3049f8 100644
--- a/iopaint/__init__.py
+++ b/iopaint/__init__.py
@@ -1,6 +1,14 @@
 import os
 
 os.environ["PYTORCH_ENABLE_MPS_FALLBACK"] = "1"
+# https://github.com/pytorch/pytorch/issues/27971#issuecomment-1768868068
+os.environ["ONEDNN_PRIMITIVE_CACHE_CAPACITY"] = "1"
+os.environ["LRU_CACHE_CAPACITY"] = "1"
+# prevent CPU memory leak when run model on GPU
+# https://github.com/pytorch/pytorch/issues/98688#issuecomment-1869288431
+# https://github.com/pytorch/pytorch/issues/108334#issuecomment-1752763633
+os.environ["TORCH_CUDNN_V8_API_LRU_CACHE_LIMIT"] = "1"
+
 
 import warnings
 
diff --git a/iopaint/batch_processing.py b/iopaint/batch_processing.py
index 2f8e082..393a720 100644
--- a/iopaint/batch_processing.py
+++ b/iopaint/batch_processing.py
@@ -1,9 +1,10 @@
 import json
-import cv2
 from pathlib import Path
 from typing import Dict, Optional
-from PIL import Image
 
+import cv2
+import psutil
+from PIL import Image
 from loguru import logger
 from rich.console import Console
 from rich.progress import (
@@ -14,10 +15,10 @@ from rich.progress import (
     TextColumn,
     BarColumn,
     TaskProgressColumn,
-    TimeRemainingColumn,
 )
 
 from iopaint.helper import pil_to_bytes
+from iopaint.model.utils import torch_gc
 from iopaint.model_manager import ModelManager
 from iopaint.schema import InpaintRequest
 
@@ -119,3 +120,8 @@ def batch_inpaint(
                 fw.write(img_bytes)
 
             progress.update(task, advance=1)
+            torch_gc()
+            # pid = psutil.Process().pid
+            # memory_info = psutil.Process(pid).memory_info()
+            # memory_in_mb = memory_info.rss / (1024 * 1024)
+            # print(f"原图大小：{img.shape},当前进程的内存占用：{memory_in_mb}MB")
diff --git a/iopaint/model_manager.py b/iopaint/model_manager.py
index 006fc0c..9be5c5d 100644
--- a/iopaint/model_manager.py
+++ b/iopaint/model_manager.py
@@ -69,6 +69,7 @@ class ModelManager:
 
         raise NotImplementedError(f"Unsupported model: {name}")
 
+    @torch.inference_mode()
     def __call__(self, image, mask, config: InpaintRequest):
         """