This package is forked from https://github.com/ivalab/grasp_multiObject

• Remember to change python path in train.py and predict.py
• change this: sys.path.insert(1, "/home/<your_pc_name>/.local/lib/python3.5/site-packages/")
• python may be different for different pc, please check your path
• Download trained model for grasp on dropbox drive
• put under <ros_ws>/src/<repo>/grasp_rcnn/output/res50/train/default/

cd <ros_ws>/src/<repo>/ source create_catkin_ws.sh cd <ros_ws> . /devel/setup.bash roslaunch realsense2_camera rs_rgbd.launch 

open new terminal

. /devel/setup.bash . ../catkin_workspace/install/setup.bash --extend cd <ros_ws>/src/<repo>/ rosrun grasp_rcnn demo_graspRGD.py 

cd <ros_ws>/src/<repo>/ source create_catkin_ws.sh cd <ros_ws> . /devel/setup.bash roslaunch realsense2_camera rs_rgbd.launch 

open new terminal

cd <ros_ws>/src/<repo> rosrun custom_stra grcnn_module.py 

This repo borrows tons of code from

• tf-faster-rcnn by endernewton
• grasp_multiObject_multiGrasp by ivalab