update README

README.md

+<!--
+ * @Author: yangxue && xue.yang@waytous.com
+ * @Date: 2025-04-07 09:39:32
+ * @LastEditors: yangxue xue.yang@waytous.com
+ * @LastEditTime: 2025-04-07 09:57:50
+ * @FilePath: /ubuntu/projects/deepinfer/README.md
+ * @Description: 
+ * 
+ * Copyright (c) 2025 by ${git_name_email}, All Rights Reserved. 
+-->
 
 
 # 1. 编译
@@ -6,14 +16,28 @@
 
 `cd DeepInfer`
 
-修改CMakeLists.txt中cuda、tensorrt、yaml、gflag和glog等库的路径
+- 根据实际情况修改CMakeLists.txt中cuda、tensorrt、yaml、gflag和glog等库的路径,。天准域控的路径如下：
+```
+# cuda
+link_directories(/usr/local/cuda/lib64)
+include_directories(/usr/local/cuda/include)
+# tensorrt
+link_directories(/usr/lib/aarch64-linux-gnu)
+include_directories(/usr/include/aarch64-linux-gnu)
+# gflags, glog
+include_directories(/usr/include/gflags)
+include_directories(/usr/include/glog)
+```
+
+
 
 `mkdir build && cd build`
 
-`cmake ..`
+`cmake ..` or `cmake -DCUDA_nppicom_LIBRARY=/usr/local/cuda/lib64/libnppim.so ..`
 
 `make -j32`
 
+(部署时，完成编译即可)
 
 # 2. 测试
 
@@ -47,5 +71,8 @@
 继承基类BaseTask，根据不同的任务，将不同模型的输入输出结合并处理
 
 
-
+# 4. 压缩部署文件
+```
+ zip -r deepinfer.zip deepinfer/ -x "deepinfer/.vscode/*" -x "deepinfer/build/*" -x "deepinfer/configs/*" -x "deepinfer/test/*" -x "deepinfer/.git/*"
+ ```