文章目录
- 1.1环境
- 1.2下载地址
- 1)操作步骤
- 2)添加环境变量
- 3)检查是否安装成功
- 2.1其它环境
1.1环境
1)Ubuntu20.042)cuda111+torch1.8.0+py38
1.2下载地址
https://developer.nvidia.com/nvidia-tensorrt-download
1)操作步骤
conda activate
cd /home/duduzai/Documents/TensorRT-8.0.1.6
pip install tensorrt-8.0.1.6-cp38-none-linux_x86_64.whl
2)添加环境变量
gedit ~/.bashrc
export LD_LIBRARY_PATH=/home/duduzai/Documents/TensorRT-8.0.1.6/lib:$LD_LIBRARY_PATH
export LD_LIBRARY_PATH=/usr/lib64:/usr/local/cuda-11.1/targets/x86_64-linux/lib:/home/duduzai/cuda+cudnn/cudnn-11.2-linux-x64-v8.1.1.33/cuda/lib64:$LD_LIBRARY_PATH
source ~/.bashrc
3)检查是否安装成功
import tensorrt as trt
print(trt.__version__)
2.1其它环境
pip install opencv-python
pip install tensorboard
pip install open3d