【B站保姆级视频教程：Jetson配置YOLOv11环境（六）PyTorchTorchvision安装】

1. 安装PyTorch

1.1安装依赖项

sudo apt install libopenblas-dev

libopenblas-dev作用：提供优化的BLAS（Basic Linear Algebra Subprograms）库，用于高效执行线性代数运算。
影响：PyTorch依赖于高效的线性代数运算来加速深度学习模型的训练和推理。libopenblas-dev提供了优化的BLAS实现，可以显著提升PyTorch的性能，尤其是在CPU上运行时。

1.2 下载torch wheel 安装包

前往PyTorch for Jetson，下载所安装的jetpack版本支持的最高版本的torch wheel 安装包到Downloads目录下。

cd /Downloads
wget https://developer.download.nvidia.cn/compute/redist/jp/v512/pytorch/torch-2.1.0a0+41361538.nv23.06-cp38-cp38-linux_aarch64.whl

例如：jetpack5.1.x对应下图中红框的torch安装包，需注意Python 版本为 3.8。

1.3 安装

pip install torch-2.1.0a0+41361538.nv23.06-cp38-cp38-linux_aarch64.whl

2. 安装torchvisiion

2.1 安装依赖

pip install numpy requests Pillow
sudo apt install libjpeg-dev libpng-dev zlib1g-dev libpython3-dev libavcodec-dev libavformat-dev libswscale-dev 

2.2 编译安装torchvision

torchvision暂未发布直接能pip安装的whl版本，因此直接从源码编译。

2.2.1 Torchvisiion版本选择

以torch2.1.0为例，对应的torchvisiion版本为0.16.x。
torch与torchvision版本对应关系

torch	torchvision	Python
main / nightly	main / nightly	>=3.9, <=3.12
2.5	0.20	>=3.9, <=3.12
2.4	0.19	>=3.8, <=3.12
2.3	0.18	>=3.8, <=3.12
2.2	0.17	>=3.8, <=3.11
2.1	0.16	>=3.8, <=3.11
2.0	0.15	>=3.8, <=3.11

2.2.2 下载torchvisiion到Downloads目录下

（1）网络ok的话，直接克隆到本地。

cd ./Downloads
git clone --branch v0.16.2 https://github.com/pytorch/vision

（2）网络不行clone慢的话，直接下载压缩包到PC

再上传jetson，解压即可

unzip vision-0.16.2.zip

2.2.3 编译安装torchvision

cd vision-0.16.2					# 进入torchvision目录
export BUILD_VERSION=0.16.2  		# 将BUILD_VERSION环境变量设置为值 0.16.2  
python3 setup.py install --user		# 使用 Python 的 setuptools 工具将vision包安装到当前用户的本地目录中

需要等待30min左右，出现以下提示则安装成功

安装成功后退出torchvision的安装目录再import torchvision进行验证，否则会出现以下warning

(pytorch) nx@nx-desktop:~/Downloads/vision-0.15.2$ python
Python 3.8.18 (default, Sep 11 2023, 13:19:25) 
[GCC 11.2.0] :: Anaconda, Inc. on linux
Type "help", "copyright", "credits" or "license" for more information.
>>> import torchvision
/home/nx/Downloads/vision-0.15.2/torchvision/io/image.py:13: UserWarning: Failed to load image Python extension: ''If you don't plan on using image functionality from `torchvision.io`, you can ignore this warning. Otherwise, there might be something wrong with your environment. Did you have `libjpeg` or `libpng` installed before building `torchvision` from source?warn(
/home/nx/Downloads/vision-0.15.2/torchvision/__init__.py:25: UserWarning: You are importing torchvision within its own root folder (/home/nx/Downloads/vision-0.15.2). This is not expected to work and may give errors. Please exit the torchvision project source and relaunch your python interpreter.warnings.warn(message.format(os.getcwd()))

2.3 安装过程可能出现的bug

若出现error: [Errno 2] No such file or directory: ':/usr/local/cuda/bin/nvcc'，请参照：

jetson编译torchvision出现 No such file or directory: ‘:/usr/local/cuda/bin/nvcc‘

3. 验证

查看pytorch运行时真正调用的cuda、cudnn版本：

python -c "import torch; import torchvision; print('PyTorch version:', torch.__version__); print('CUDA available:', torch.cuda.is_available()); print('CUDA version:', torch.version.cuda); print('cuDNN enabled:', torch.backends.cudnn.enabled); print('cuDNN version:', torch.backends.cudnn.version()); print('Torchvision version:', torchvision.__version__)"

---

2025-2-4 19:47:09更新：发现了更便捷的安装方式，ultralytics官方已经提供了编译好的torchvision whl包。

4. 更便捷的安装方式

在安装完torch和torchvision的依赖之后，运行下述两行命令即可。若下载慢，出现WARNING: Retrying，直接先下载到本地PC再上传jetson安装即可。

4.1 JetPack 5.1.2

安装 torch 2.1.0 和 torchvision 0.16.2

pip install https://github.com/ultralytics/assets/releases/download/v0.0.0/torch-2.1.0a0+41361538.nv23.06-cp38-cp38-linux_aarch64.whl
pip install https://github.com/ultralytics/assets/releases/download/v0.0.0/torchvision-0.16.2+c6f3977-cp38-cp38-linux_aarch64.whl

4.2 JetPack 6.1

安装 torch 2.5.0 和 torchvision 0.20

pip install https://github.com/ultralytics/assets/releases/download/v0.0.0/torch-2.5.0a0+872d972e41.nv24.08-cp310-cp310-linux_aarch64.whl
pip install https://github.com/ultralytics/assets/releases/download/v0.0.0/torchvision-0.20.0a0+afc54f7-cp310-cp310-linux_aarch64.whl

安装 cuSPARSELt 的依赖性问题 torch 2.5.0

wget https://developer.download.nvidia.com/compute/cuda/repos/ubuntu2204/arm64/cuda-keyring_1.1-1_all.deb
sudo dpkg -i cuda-keyring_1.1-1_all.deb
sudo apt-get update
sudo apt-get -y install libcusparselt0 libcusparselt-dev