This repository contains the official implementation of the following paper:
(IROS 2024) Learning Generalizable Tool-use Skills through Trajectory Generation
Carl Qi*, Yilin Wu*, Lifan Yu, Haoyue Liu, Bowen Jiang, Xingyu Lin**, David Held**
-
Initialize a conda environment (python=3.6) and install
python3 -m pip install -e . -
Install torch (version 1.9.0 tested)
- We tested
pip install torch==1.9.0+cu111 torchvision==0.10.0+cu111 torchaudio==0.9.0 -f https://download.pytorch.org/whl/torch_stable.htmlon RTX 3090.
- We tested
-
Install packages for computing the EMD loss:
- pykeops (1.5) by
running
pip install pykeops==1.5 - geomloss by running
pip install geomloss
- pykeops (1.5) by
running
-
Install necessary packages for PointFlow
-
(optional) Install chester from https://github.com/Xingyu-Lin/chester.
-
The training code for ToolGen has 2 parts:
- The first part is training the PointFlow model. The training script is in
PointFlow/scripts/gen_multitool.sh - The second part is training the trajectory model. The training script code is in
core/toolgen/train_model.py, and the model architecture and inference code is incore/toolgen/bc_agent.py.
- The first part is training the PointFlow model. The training script is in
-
The launch script that leverages chester to train/evaluate the ToolGen trajectory model is under
core/toolgen/launchers/launch_train_bc.py. Alternatively, one can write a custom script calling run_task intrain_model.pywithout using chester.
The training data from the ToolGen paper will be released soon, stay tuned!
If you find this codebase useful in your research, please consider citing:
@INPROCEEDINGS{qi2024toolgen,
author={Qi, Carl and Wu, Yilin and Yu, Lifan and Liu, Haoyue and Jiang, Bowen and Lin, Xingyu and Held, David},
booktitle={2024 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS)},
title={Learning Generalizable Tool-use Skills through Trajectory Generation},
year={2024},
volume={},
number={},
pages={2847-2854},
keywords={Point cloud compression;Deformable models;Shape;Autonomous systems;Affordances;Data models;Cleaning;Trajectory;Intelligent robots},
doi={10.1109/IROS58592.2024.10801653}}
(CoRL 2022) Planning with Spatial-Temporal Abstraction from Point Clouds for Deformable Object Manipulation
Xingyu Lin*, Carl Qi*, Yunchu Zhang, Zhiao Huang, Katerina Fragkiadaki, Yunzhu Li, Chuang Gan, David Held
(RA-L 2022) Learning Closed-loop Dough Manipulation Using a Differentiable Reset Module
Carl Qi, Xingyu Lin, David Held