Zser#132
Open
zhousheng-zser wants to merge 41 commits into
Open
Conversation
## 一、硬件环境 ### 1.1 硬件 #### GPU(图形处理器) - **架构**: NVIDIA GPU,支持 CUDA 计算能力 7.0 及以上 - **推荐平台**: - **NVIDIA Jetson Orin** (官方测试平台,推荐) - **显存**: 建议至少 **8GB** 以上(推荐 16GB+) - **计算能力**: SM 7.0+ (Volta/Turing/Ampere/Ada Lovelace 架构) #### CPU(中央处理器) - **架构**: - **ARM64 (aarch64)** - 用于 Jetson 平台(推荐) - **核心数**: 建议 4 核及以上 - **内存**: 建议至少 **16GB** RAM(推荐 32GB+) ### 1.2 可选硬件(根据应用场景) #### 激光雷达(如果使用实时检测) - **支持的雷达**: - Hesai 系列(需要 HesaiLidar_SDK_2.0) - Innovusion 系列(需要 inno-lidar SDK) - **接口**: 以太网接口(用于接收点云数据) #### 网络 - 用于接收点云数据或提供 HTTP 服务 - 建议千兆以太网接口 --- ## 二、软件环境 ### 2.1 操作系统 #### 主机系统 - **操作系统**: Ubuntu 22.04.5 LTS (Jammy Jellyfish) - **系统架构**: ARM64 (aarch64) #### 容器内系统 - **操作系统**: Ubuntu 22.04.4 LTS (Jammy Jellyfish) - **系统架构**: ARM64 (aarch64) ### 2.2 NVIDIA 驱动和 CUDA #### NVIDIA 驱动 - **版本**: 540.4.0 - **CUDA 驱动版本**: 12.6(驱动支持的最高 CUDA 版本) #### CUDA Toolkit - **版本**: CUDA 11.8.89 - **构建信息**: Built on Wed_Sep_21_10:43:33_PDT_2022 - **编译器版本**: cuda_11.8.r11.8/compiler.31833905_0 - **安装路径**: `/usr/local/cuda-11.8/` - **环境变量配置**: - `CUDA_Lib=/usr/local/cuda-11.8/targets/aarch64-linux/lib/` - `CUDA_Inc=/usr/local/cuda-11.8/targets/aarch64-linux/include/` - `CUDA_Bin=/usr/local/cuda-11.8/bin/` - `CUDA_HOME=/usr/local/cuda-11.8/` #### cuDNN - **版本**: cuDNN 8.9.4.25 (for CUDA 12.2) - **安装路径**: `/lib/aarch64-linux-gnu/` - **环境变量配置**: `CUDNN_Lib=/lib/aarch64-linux-gnu/` - **已安装包**: - `libcudnn8` (8.9.4.25-1+cuda12.2) - `libcudnn8-dev` (8.9.4.25-1+cuda12.2) ### 2.3 TensorRT - **版本**: TensorRT 8.6.2.3 (trtexec 显示 v8602) - **安装路径**: - 库文件: `/lib/aarch64-linux-gnu/` - 头文件: `/usr/include/aarch64-linux-gnu/` - 工具: `/usr/src/tensorrt/bin/` - **环境变量配置**: - `TensorRT_Lib=/lib/aarch64-linux-gnu/` - `TensorRT_Inc=/usr/include/aarch64-linux-gnu/` - `TensorRT_Bin=/usr/src/tensorrt/bin/` - **已安装包**: - `libnvinfer-bin` (8.6.2.3-1+cuda12.2) - `libnvinfer-dev` (8.6.2.3-1+cuda12.2) - `libnvinfer-plugin8` (8.6.2.3-1+cuda12.2) - `libnvinfer-dispatch8` (8.6.2.3-1+cuda12.2) - 以及其他 TensorRT 相关开发库 - **必需工具**: `trtexec` (位于 `/usr/src/tensorrt/bin/trtexec`) ### 2.4 编译工具链 #### C/C++ 编译器 - **GCC**: 11.4.0 (Ubuntu 11.4.0-1ubuntu1~22.04) - **G++**: 11.4.0 (Ubuntu 11.4.0-1ubuntu1~22.04) - **CMake**: 3.22.5 - **编译标准**: C++17 #### 构建工具 - **Make**: 标准构建工具 - **Git**: 用于克隆代码仓库 - **Git LFS**: 用于下载大文件(模型权重等) ### 2.5 Python 环境(用于模型导出和训练) #### Python 版本 - **Python**: 3.10.12 #### Python 包(当前已安装版本) ```bash torch==1.14.0a0+44dac51c.nv23.2 # PyTorch (NVIDIA 定制版本) torchvision==0.13.0 onnx==1.17.0 # ONNX 模型格式支持 onnxsim==0.4.36 # ONNX 模型简化 (onnx-simplifier) onnx-graphsurgeon==0.3.27 # ONNX 图操作 onnxruntime==1.19.2 spconv==2.3.6 # 稀疏卷积库 (本地编译版本) numpy==1.23.5 numba==0.58.0 opencv-python==4.11.0.86 scikit-image==0.19.3 ``` ### 2.6 Docker(当前部署方式) #### Docker 版本 - **Docker**: 27.5.1 (build 27.5.1-0ubuntu3~22.04.2) - **Docker 运行时**: nvidia (用于 GPU 访问) - **容器镜像**: `my-cuda-env:251023-release` (约 39.9GB) #### Docker 运行配置 - **网络模式**: host - **特权模式**: 启用 (--privileged) - **PID 模式**: host - **IPC 模式**: host - **挂载点**: - `/etc/localtime:/etc/localtime:ro` - `/home/nvidia/zser:/home/zser` --- ## 三、部署到其他电脑需要的信息 #### 环境变量(需要根据目标系统调整) ```bash # TensorRT 路径 export TensorRT_Lib=/lib/aarch64-linux-gnu/ # 或 /usr/lib/x86_64-linux-gnu/ export TensorRT_Inc=/usr/include/aarch64-linux-gnu/ # 或 /usr/include/ export TensorRT_Bin=/usr/src/tensorrt/bin/ # CUDA 路径 export CUDA_Lib=/usr/local/cuda-11.8/targets/aarch64-linux/lib/ export CUDA_Inc=/usr/local/cuda-11.8/targets/aarch64-linux/include/ export CUDA_Bin=/usr/local/cuda-11.8/bin/ export CUDA_HOME=/usr/local/cuda-11.8/ # cuDNN 路径 export CUDNN_Lib=/lib/aarch64-linux-gnu/ # 或 /usr/lib/x86_64-linux-gnu/ # CUDA 计算能力(SM) export CUDASM=87 # 当前系统:Jetson Orin (SM 8.7) ``` #### 编译配置 - **当前系统 CUDA 计算能力**: SM 8.7 (Jetson Orin) - **其他 GPU 的 SM 版本参考**: - Jetson Orin: SM 8.7 (当前系统) - RTX 30 系列: SM 8.6 - RTX 40 系列: SM 8.9 - A100: SM 8.0 ---
2. web网页画线代码
2. 切换一些路径
2. 新增停止线
Author
|
合并到主分支 |
2. 新环境配置, 3. 点云等距线 强度=-4
2. 删除httplib.h 3. 调整下追踪器
2.更新模型,连带修改配置文件 3.减小nms,防车叠车
2. 配置微调
2.没检测到车再给一次机会
2.新增速度校准配置和速度获取线配置. 3.速度获取线画图.
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
No description provided.