Learning Grasping with Self-Supervised Deep Reinforcement Learning

This repo is final project for 2021 Deep Learning and Practice class

teacher : 陳永昇、吳毅成、彭文孝

---

clone repo

     $ git clone --recursive [email protected]:kuolunwang/Grasping-with-SSDRL.git

You should run this code on a GPU computer

---

Building docker image

     $ cd Docker && source build.sh

How to run

    $ cd Docker && source docker_run.sh
    Docker $ cd Grasping-with-SSDRL && source catkin_make.sh
    Docker $ source environment.sh

If you want to enter same container

    $ cd Docker && source docker_join.sh
    Docker $ cd Grasping-with-SSDRL && source environment.sh

Online learning

     $ roslaunch arm_bringup ur5.launch
     $ rosrun algorithm main.py