lsbnb/cuda repository overview

Edited CUDA docker image from nvidia/cuda⁠

⁠NVIDIA License Agreements (from https://hub.docker.com/r/nvidia/cuda⁠)

By downloading these images, you agree to the terms of the license agreements for NVIDIA software included in the images.

⁠CUDA Toolkit License

To view the license for the CUDA Toolkit included in this image, click here⁠

⁠cuDNN Toolkit

To view the license for the cuDNN Toolkit included in this image, click here⁠

⁠Tag:

• v1: Edit from nvidia/cuda:9.0-cudnn7-runtime and install nano, htop, tensorflow, pytorch...etc
• 10.1: Edit from nvidia/cuda:10.1-cudnn7-runtime-ubuntu16.04 and install nano, htop...etc

⁠Requirements:

• OS: Ubuntu 16.0.4
• nvidia-docker >= 1.0.1-1
• NVIDIA GPU's Gencode = SM61 (NVIDIA GPU:GTX 1080, GTX 1070, GTX 1060, GTX 1050, GTX 1030, Titan Xp, Tesla P40, Tesla P4)
• CUDA = 9.0 for tag v1
• CUDA = 10.1 for tag 10.1

⁠Notes:

Need to use nvidia-docker before running the application.Link:https://github.com/NVIDIA/nvidia-docker⁠ NVIDIA GPU must meet the requirements

⁠Installation with selected tag:

nvidia-docker pull lsbnb/cuda:<tag>

⁠Example (Using tag: v1):

nvidia-docker pull lsbnb/cuda:v1

nvidia-docker run -p 8888:8888 -ti lsbnb/cuda:v1

⁠Caution:

To use it with GPU, you MUST run like the following instruction:

nvidia-docker run -p 8888:8888 -ti lsbnb/cuda:v1

NOT:

docker run -p 8888:8888 -ti lsbnb/cuda:v1