Skip to content

Latest commit

 

History

History
94 lines (63 loc) · 2.12 KB

File metadata and controls

94 lines (63 loc) · 2.12 KB

Installation

Python

Create a virtual environment with venv

sudo apt install python3.13-venv
python3.13 -m venv .venv
. .venv/bin/activate

Validate python version in the venv

(.venv) $ python -V
Python 3.13.5

Install following dependencies

1. slobot

git clone git+https://github.com/alexis779/slobot.git
cd slobot
pip install -e .

2. Robot Configuration

Ensure the robot configuration in available in slobot.config package.

cd ..
git clone https://github.com/google-deepmind/mujoco_menagerie
cd slobot
ln -s ../mujoco_menagerie/trs_so_arm100 src/slobot/config/trs_so_arm100

3. LeRobot

pip install git+https://github.com/huggingface/lerobot.git

4. Genesis

pip install git+https://github.com/Genesis-Embodied-AI/Genesis.git

Also refer to the installation guide. Make sure to run the hello world example successfully.

Known issue

On Ubuntu, Qt5 library may be incompatible with pymeshlab native library. See reported issue. As a workaround, give precedence to the python module QT library instead of the Ubuntu system QT library.

SITE_PACKAGES=`pip show pymeshlab | grep Location | sed 's|Location: ||'`
PYMESHLAB_LIB=$SITE_PACKAGES/pymeshlab/lib

Make sure the symbol is found

strings $PYMESHLAB_LIB/libQt5Core.so.5 | grep _ZdlPvm

Finally, configure LD_LIBRARY_PATH to overwrite QT library path,

LD_LIBRARY_PATH=$PYMESHLAB_LIB python <script.py>

Docker

Local

Build docker image:

docker build -f docker/Dockerfile.local -t slobot .

Run docker container. Make sure to enable DRI for hardware graphics acceleration.

docker run -it --security-opt no-new-privileges=true -p 7860:7860 --device=/dev/dri -v $PWD:/home/user/app slobot