简体   繁体   English

CUDA libcublas.so.11 在 ONNX Docker 容器内使用 GPU 时出错

[英]CUDA libcublas.so.11 Error when using GPUs inside ONNX Docker Container

While programming with python3.6 on DGX Station (NVIDIA) based on ONNX runtime environment;在基于ONNX运行环境的DGX Station(NVIDIA)上使用python3.6编程时;

Using following libraries;使用以下库;

mxnet==1.5.x onnxruntime-gpu==1.7.x mxnet==1.5.x onnxruntime-gpu==1.7.x

I see following error我看到以下错误

OSError: libcublas.so.11: cannot open shared object file: No such file or directory

Fixed it by updating CUDA to 11.0通过将 CUDA 更新为 11.0 来修复它

apt update
apt install cuda-11.0
pip3 uninstall mxnet //and reinstall latest version
pip3 uninstall onnxruntime-gpu //and resinstall latest version

then a new error started appearing然后一个新的错误开始出现

onnxruntime cuda failure 100 no cuda-capable device is detected

then it was noted that docker was not started with nvidia runtime so following was added to the docker-compose file然后注意到 docker 未使用 nvidia 运行时启动,因此将以下内容添加到 docker-compose 文件中

runtime: nvidia

and all started working一切都开始工作了

暂无
暂无

声明:本站的技术帖子网页,遵循CC BY-SA 4.0协议,如果您需要转载,请注明本站网址或者原文地址。任何问题请咨询:yoyou2525@163.com.

相关问题 仅在使用GUI时Tensorflow的“ImportError:libcublas.so.10.0” - “ImportError: libcublas.so.10.0” for Tensorflow ONLY when using GUI 张量流的libcublas.so.8.0错误 - libcublas.so.8.0 error with tensorflow Tensorflow 只显示“在本地成功打开 CUDA 库 libcublas.so.10.0”,而没有关于 cudnn - Tensorflow shows only "successfully opened CUDA library libcublas.so.10.0 locally" and nothing about cudnn 使用CUDA8.0和cuDNN5.1安装tensorflow1.2时显示'ImportError:libcublas.so.9.0' - Install tensorflow1.2 with CUDA8.0 and cuDNN5.1 shows 'ImportError: libcublas.so.9.0' 在Docker容器中运行python脚本时导入错误? - Import error when running python script inside docker container? 无法编译cuda_ndarray.cu:libcublas.so.7.5:无法打开共享对象文件 - Failed to compile cuda_ndarray.cu: libcublas.so.7.5: cannot open shared object file Pycharm 使用带 GPU 的 docker 进行调试 - Pycharm debugging using docker with GPUs 在 docker 中运行 CuFFT 时出现 CUDA cudaErrorInsufficientDriver 错误 - CUDA cudaErrorInsufficientDriver error when running an CuFFT in docker Theano for Keras:无法编译cuda_ndarray.cu:libcublas.so.7.5:无法打开共享对象文件:无此类文件或目录 - Theano for Keras: Failed to compile cuda_ndarray.cu: libcublas.so.7.5: cannot open shared object file: No such file or directory 使用 docker-compose 启动 Jupyter notebook 时激活 Docker 容器内的 conda env - Activating a conda env inside a Docker container when using docker-compose to start Jupyter notebook
 
粤ICP备18138465号  © 2020-2024 STACKOOM.COM