/tiago_docker_tensorflow

CUDA and CUDNN supported ROS development environment

Primary LanguageShell

TIAGO-DOCKER-TENSORFLOW

Original container: https://github.com/nozomisk/tiago-docker

The Dockerfile was modified to meed the requirements of the project ros_contact_graspnet.

ROS Version: Melodic

Requirements

This container is meant to run with a capable GPU. It has been tested with NVIDIA RTX 2060 and NVIDIA RTX 3060.

Docker. Docker-compose. Nvidia docker. CUDA and CUDNN.

How to run the container

  1. To build the container run sh install.sh
  2. Export the following environment variables to your shell session: export TIAGO_DOCKER_ROOT=<path-to-this-repository> export PATH=$PATH:$TIAGO_DOCKER_ROOT/bin It is recommended that these variables are appended to your ~/.bashrc file to automate this process.
  3. Run tiago_up to start the container
  4. Run tiago_bash to execute a shell session within the container.
  5. Run tiago_stop to stop the container.
  6. Run tiago_down to remove the container created by tiago_up.

Hints

The development workspace of the container is within the tiago_home directory created after the building process.

If you want to create different development enviroments with this container it is recommended that you change follow the same installation process but change the name of the environment variables that were set earlier, and the name of the scripts under the ./bin directory.

If you would like to use this development environment with PyCharm, then please refer to https://github.com/nozomisk/tiago-docker