This is the code for the paper: "Latent Space Roadmap for Visual Action Planning of Deformable and Rigid Object Manipulation"
Visit the dedicated website: Latent Space Roadmap website for more information.
If you use this code in your work, please cite it as follows:
@inproceedings{lippi2020latent,
title={Latent Space Roadmap for Visual Action Planning of Deformable and Rigid Object Manipulation},
author={Lippi, Martina and Poklukar, Petra and Welle, Michael C and Varava, Anastasiia and Yin, Hang and Marino, Alessandro and Kragic, Danica},
booktitle={IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS)},
year={2020}
}
pip install -r requirements.txt
Download LSR stacking datasets:
cd datasets/
python get_datasets.py
cd ..
To make train and test split use:
python preprocess_dataset.py
To train the VAE use:
python train_VAE.py --exp_vae=VAE_UnityStacking_L1 --cuda=True
To train the APN use:
touch models/APN_UnityStacking_evaluation_results.txt
python train_APN_stacking.py \
--exp_apn=APN_UnityStacking_L1 \
--seed=98765 \
--generate_new_splits=1 \
--generate_apn_data=1 \
--train_apn=1 \
--eval_apn=1 \
--cuda=True
python lsr_stacking_example.py --seed=98765 --lable_ls=True --build_lsr=True --example=True
You should get a image like this in the root folder as a result: (depending on your random seed)
