The gqcnn Python package is for training and analysis of Grasp Quality Convolutional Neural Networks (GQ-CNNs). It is part of the ongoing Dexterity-Network (Dex-Net) project created and maintained by the AUTOLAB at UC Berkeley.

Please see the docs for installation and usage instructions.

If you use any part of this code in a publication, please cite the appropriate Dex-Net publication.

(WARNING) Download pre-trained model from our synology. Please see details on bellow information.

Download code on your catkin_ws.

git clone -b melodic-devel --single-branch https://github.com/rise-lab-skku/rise-gqcnn.git

Recommanded: Use virtual environment and activate it.

cd rise-gqcnn
virtualenv -p python3.6 --system-site-packages venv
source venv/bin/activate

Change directories into the gqcnn repository and run the pip installation.

pip install .

(WARNING) Official download link is broken. Please follow the bellow intruction.

Create directory in the gqcnn repository.

mkdir -p models

Download pre-trained models on models directory. The models can be found on our synology /Research Projects/2020_지능증강/NN_models/official-gqcnn-models.

Unzip pre-trained models.

cd models
unzip -a GQCNN-4.0-PJ.zip
unzip -a GQCNN-4.0-SUCTION.zip
unzip -a FC-GQCNN-4.0-PJ.zip
unzip -a FC-GQCNN-4.0-SUCTION.zip
cd ..

Start the grasp planning service:

roslaunch gqcnn grasp_planning_service.launch ns:=pj_gqcnn model_name:=FC-GQCNN-4.0-PJ fully_conv:=true

The example ROS policy can then be queried on saved images using:

python examples/policy_ros.py --depth_image data/examples/clutter/phoxi/fcgqcnn/depth_0.npy --segmask data/examples/clutter/phoxi/fcgqcnn/segmask_0.png --camera_intr data/calib/phoxi/phoxi.intr --namespace pj_gqcnn