Dataset Card for GEMBench dataset

💎 GEneralizable vision-language robotic Manipulation Benchmark Dataset

A benchmark to systematically evaluate generalization capabilities of vision-and-language robotic manipulation policies. Built upon the RLBench simulator.

💻 GEMBench Project Webpage: https://www.di.ens.fr/willow/research/gembench/

📈 Leaderboard: https://paperswithcode.com/sota/robot-manipulation-generalization-on-gembench

Dataset Structure

Dataset structure is as follows:

  - gembench
      - train_dataset
          - microsteps: 567M, initial configurations for each episode
          - keysteps_bbox: 160G, extracted keysteps data
          - keysteps_bbox_pcd: (used to train 3D-LOTUS)
              - voxel1cm: 10G, processed point clouds
              - instr_embeds_clip.npy: instructions encoded by CLIP text encoder
          - motion_keysteps_bbox_pcd: (used to train 3D-LOTUS++ motion planner)
              - voxel1cm: 2.8G, processed point clouds
              - action_embeds_clip.npy: action names encoded by CLIP text encoder
      - val_dataset
          - microsteps: 110M, initial configurations for each episode
          - keysteps_bbox_pcd:
              - voxel1cm: 941M, processed point clouds
      - test_dataset
          - microsteps: 2.2G, initial configurations for each episode

🛠️ Benchmark Installation

1. Create and activate your conda environment:

conda create -n gembench python==3.10

conda activate gembench

2. Install RLBench Download CoppeliaSim (see instructions here)

# change the version if necessary
wget https://www.coppeliarobotics.com/files/V4_1_0/CoppeliaSim_Edu_V4_1_0_Ubuntu20_04.tar.xz
tar -xvf CoppeliaSim_Edu_V4_1_0_Ubuntu20_04.tar.xz

Add the following to your ~/.bashrc file:

export COPPELIASIM_ROOT=$(pwd)/CoppeliaSim_Edu_V4_1_0_Ubuntu20_04
export LD_LIBRARY_PATH=$LD_LIBRARY_PATH:$COPPELIASIM_ROOT
export QT_QPA_PLATFORM_PLUGIN_PATH=$COPPELIASIM_ROOT

Install Pyrep and RLBench

git clone https://github.com/cshizhe/PyRep.git
cd PyRep
pip install -r requirements.txt
pip install .
cd ..

# Our modified version of RLBench to support new tasks in GemBench
git clone https://github.com/rjgpinel/RLBench
cd RLBench
pip install -r requirements.txt
pip install .
cd ..

Evaluation

Please, check 3D-LOTUS++ official code repository for evaluation:

https://github.com/vlc-robot/robot-3dlotus?tab=readme-ov-file#evaluation

Citation

If you use our GemBench benchmark or find our code helpful, please kindly cite our work:

BibTeX:

 @inproceedings{garcia25gembench,
    author    = {Ricardo Garcia and Shizhe Chen and Cordelia Schmid},
    title     = {Towards Generalizable Vision-Language Robotic Manipulation: A Benchmark and LLM-guided 3D Policy},
    booktitle = {IEEE International Conference on Robotics and Automation (ICRA)},
    year      = {2025}
}

Contact

Ricardo Garcia-Pinel