diff --git a/docs/README_Windows_CUDA_Acceleration_en_US.md b/docs/README_Windows_CUDA_Acceleration_en_US.md
index e8252145..7aa96990 100644
--- a/docs/README_Windows_CUDA_Acceleration_en_US.md
+++ b/docs/README_Windows_CUDA_Acceleration_en_US.md
@@ -65,7 +65,7 @@ If your graphics card has at least 8GB of VRAM, follow these steps to test CUDA-
 1. **Overwrite the installation of torch and torchvision** supporting CUDA.
 
    ```
-   pip install --force-reinstall torch torchvision --index-url https://download.pytorch.org/whl/cu118
+   pip install --force-reinstall torch==2.3.1 torchvision==0.18.1 --index-url https://download.pytorch.org/whl/cu118
    ```
 
 2. **Modify the value of `"device-mode"`** in the `magic-pdf.json` configuration file located in your user directory.
diff --git a/docs/README_Windows_CUDA_Acceleration_zh_CN.md b/docs/README_Windows_CUDA_Acceleration_zh_CN.md
index 0b156994..dcdb71ee 100644
--- a/docs/README_Windows_CUDA_Acceleration_zh_CN.md
+++ b/docs/README_Windows_CUDA_Acceleration_zh_CN.md
@@ -66,7 +66,7 @@ pip install -U magic-pdf[full] --extra-index-url https://wheels.myhloli.com -i h
 **1.覆盖安装支持cuda的torch和torchvision**
 
 ```bash
-pip install --force-reinstall torch torchvision --index-url https://download.pytorch.org/whl/cu118
+pip install --force-reinstall torch==2.3.1 torchvision==0.18.1 --index-url https://download.pytorch.org/whl/cu118
 ```
 
 **2.修改【用户目录】中配置文件magic-pdf.json中"device-mode"的值**
diff --git a/setup.py b/setup.py
index a1f84a0e..aaa4c1b3 100644
--- a/setup.py
+++ b/setup.py
@@ -36,7 +36,9 @@ if __name__ == '__main__':
                      "paddlepaddle==3.0.0b1;platform_system=='Linux'",
                      "paddlepaddle==2.6.1;platform_system=='Windows' or platform_system=='Darwin'",
                      ],
-            "full": ["unimernet==0.2.2",  # unimernet升级0.2.1
+            "full": ["unimernet==0.2.2",  # unimernet升级0.2.2,移除torchtext的依赖
+                     "torch>=2.2.2,<=2.3.1",  # torch2.4.0及之后版本未测试，先卡住版本上限
+                     "torchvision>=0.17.2,<=0.18.1",  # torchvision 受torch版本约束
                      "matplotlib<=3.9.0;platform_system=='Windows'",  # 3.9.1及之后不提供windows的预编译包，避免一些没有编译环境的windows设备安装失败
                      "matplotlib;platform_system=='Linux' or platform_system=='Darwin'",  # linux 和 macos 不应限制matplotlib的最高版本，以避免无法更新导致的一些bug
                      "ultralytics>=8.3.48",  # yolov8,公式检测