---
title: "Isaac Gym — paper by Makoviychuk et al. (2021)"
description: "Users build scenes from URDF or MJCF robot models and replicate each environment thousands of… 39 references, cited by 151 papers here."
authors: ["Viktor Makoviychuk", "Lukasz Wawrzyniak", "Yunrong Guo", "Michelle Lu", "Kier Storey", "Miles Macklin", "David Hoeller", "Nikita Rudin", "Arthur Allshire", "Ankur Handa", "Gavriel State"]
year: 2021
venue: "arXiv:2108.10470"
source: "https://arxiv.org/abs/2108.10470"
doi: "10.48550/arxiv.2108.10470"
arxiv: "2108.10470"
licence: "CC BY 4.0 (arXiv)"
licence_url: "http://creativecommons.org/licenses/by/4.0/"
summary_licence: "Elamur Cosmic License"
canonical: "https://research.elamur.ai/free/en/papers/makoviychuk-2021-isaac-gym-high-performance-gpu/"
---
[Elamur Research](https://md.research.elamur.ai/free/en/) · [this page in HTML](https://research.elamur.ai/free/en/papers/makoviychuk-2021-isaac-gym-high-performance-gpu/)

[home](https://md.research.elamur.ai/free/en/) / [papers](https://md.research.elamur.ai/free/en/papers/) / [2021](https://md.research.elamur.ai/free/en/papers/2021/page/2/) / Isaac Gym: High Performance GPU-Based Physics Simulation For Robot Learning

# Isaac Gym: High Performance GPU-Based Physics Simulation For Robot Learning

**Authors:** Viktor Makoviychuk, Lukasz Wawrzyniak, Yunrong Guo, Michelle Lu, Kier Storey, Miles Macklin, David Hoeller, Nikita Rudin, Arthur Allshire, Ankur Handa and Gavriel State · **Year:** 2021 · **Venue:** arXiv:2108.10470 · **Licence:** [CC BY 4.0 (arXiv)](http://creativecommons.org/licenses/by/4.0/)

**Links:** [Publisher's page](https://arxiv.org/abs/2108.10470) · [PDF](https://arxiv.org/pdf/2108.10470) · [DOI 10.48550/arxiv.2108.10470](https://doi.org/10.48550/arxiv.2108.10470) · [arXiv 2108.10470](https://arxiv.org/abs/2108.10470) · [OpenAlex](https://openalex.org/W3193576605)

**Citations:** 58 citations (OpenAlex, 2026-10-14) · [151](https://md.research.elamur.ai/free/en/papers/makoviychuk-2021-isaac-gym-high-performance-gpu/cited-by/) in our library, of the 149,439 papers whose bibliographies we read

## Summary
Reinforcement learning for robots usually pairs CPU-based physics simulators with GPU-based neural networks. That split forces constant data transfers and needs huge CPU clusters for hard tasks. Isaac Gym, from NVIDIA, removes the split by running everything on a single GPU: physics simulation with PhysX, observation and reward computation, and policy training. A Python Tensor API exposes simulator state and control buffers directly as PyTorch tensors, so data never travels through the CPU. Users build scenes from URDF or MJCF robot models and replicate each environment thousands of times, with optional per-copy variation. The authors benchmark eight tasks, from simple locomotion to dexterous in-hand manipulation, and report training that is 100 to 1,000 times faster than conventional setups. They also reproduce OpenAI's Shadow Hand cube-reorientation results in a fraction of the time and transfer policies to real ANYmal and TriFinger robots. This matters because experiments that once needed a data center can now run on one desktop GPU.

## Key points
- Eight benchmark environments: locomotion with Ant, Humanoid and ANYmal; the Ingenuity helicopter; stacking cubes with a Franka arm; and manipulation with the Shadow Hand, Allegro Hand and TriFinger. Humanoid character animation with AMP is also included.
- Reported times on one NVIDIA A100:
  - Ant locomotion: about 20 s
  - Humanoid: about 4 min
  - ANYmal: under 2 min
  - AMP animation: about 6 min
  - Shadow Hand cube rotation, standard setting: about 35 min
- OpenAI Shadow Hand reproduction, with asymmetric actor-critic, domain randomization and 0.4 rad success tolerance:
  - Feed-forward policy: 20 consecutive successes in about 1 h, versus 30 h for OpenAI.
  - LSTM policy: 37 consecutive successes in about 6 h on average, or 2.5 h for the best seed, versus about 17 h for OpenAI.
  - OpenAI's hardware was 384 CPU machines (6144 cores) plus 8 V100 GPUs running MuJoCo.
- Sim-to-real: the TriFinger cube-reposing policy reached about 55% mean success on the real robot, which was operated remotely. ANYmal locomotion also transferred to hardware.
- A basic PPO implementation and a simple task system ship with the platform; other RL libraries can be swapped in.

## Interesting ideas
- **Keep the whole loop on one device.** Simulation, reward logic and learning all live in GPU memory, which removes the transfer bottleneck. Any pipeline that alternates between CPU preprocessing and GPU models, such as data augmentation or feature extraction, could gain from the same principle.
- **Expose internal buffers as framework-native tensors.** Wrapping simulator memory as PyTorch tensors avoids copies and makes the simulator easy to plug into existing training code. The same zero-copy approach suits other backends that feed ML models.
- **Asymmetric actor-critic.** The critic sees privileged simulation information, while the actor uses only inputs that are available in reality. This lets you train with rich signals and still deploy under realistic constraints, much like using gold features only at training time.
- **Mass replication with controlled variation.** One environment is cloned thousands of times, each copy randomized slightly. This yields parallelism and robustness together, an idea that carries over to synthetic data generation.
- **Compare honestly against single-seed baselines.** The authors report both their average over seeds and their best seed, because the baseline reported only one run. Stating the seed protocol explicitly is a good habit for any benchmark comparison.

*Our summary of the paper, in our own words — not its text. The work belongs to its authors; read it at the links above.*

## Institutions

- [University of Toronto](https://md.research.elamur.ai/free/en/universities/university-of-toronto/) — University, Toronto, Canada · 110 works · 179,062 citations
- [ETH Zurich](https://md.research.elamur.ai/free/en/universities/eth-zurich/) — University, Zurich, Switzerland · 84 works · 9,181 citations
- [Nvidia](https://md.research.elamur.ai/free/en/companies/nvidia/) — Company, Santa Clara, United States · 37 works · 12,656 citations
- [Nvidia (United Kingdom)](https://md.research.elamur.ai/free/en/companies/nvidia-united-kingdom/) — Company, Reading, United Kingdom · 27 works · 11,141 citations

📚 Shelves (5): [🖥️ Computer Science](https://md.research.elamur.ai/free/en/shelves/computer-science/) · 18,614 works · [🤖 Artificial Intelligence](https://md.research.elamur.ai/free/en/shelves/computer-science/artificial-intelligence/) · 17,965 works · [🔌 Hardware and Architecture](https://md.research.elamur.ai/free/en/shelves/computer-science/hardware-and-architecture/) · 167 works · [⚙️ Engineering](https://md.research.elamur.ai/free/en/shelves/engineering/) · 188 works · [⚙️ Control and Systems Engineering](https://md.research.elamur.ai/free/en/shelves/engineering/control-and-systems-engineering/) · 62 works

🧭 Topics (1): [⚡ GPU & parallel computing](https://md.research.elamur.ai/free/en/topics/gpu-parallel-computing/) · 41 works

## Cite

Cite as: Viktor Makoviychuk, Lukasz Wawrzyniak, Yunrong Guo, Michelle Lu, Kier Storey, Miles Macklin, David Hoeller, Nikita Rudin, Arthur Allshire, Ankur Handa and Gavriel State. 2021. Isaac Gym: High Performance GPU-Based Physics Simulation For Robot Learning. arXiv:2108.10470.

The paper itself, not this page. BibTeX from the library's record of the work.

```bibtex
@misc{makoviychuk2021isaac,
  title = {{Isaac Gym: High Performance GPU-Based Physics Simulation For Robot Learning}},
  author = {Viktor Makoviychuk, Lukasz Wawrzyniak, Yunrong Guo, Michelle Lu, Kier Storey, Miles Macklin, David Hoeller, Nikita Rudin, Arthur Allshire, Ankur Handa and Gavriel State},
  year = {2021},
  doi = {10.48550/arxiv.2108.10470},
  eprint = {2108.10470},
  archivePrefix = {arXiv},
  url = {https://doi.org/10.48550/arxiv.2108.10470}
}
```

- [← previous XOR QA: Cross-lingual Open-Retrieval Question Answering](https://md.research.elamur.ai/free/en/papers/asai-2021-xor-qa-cross-lingual-open/)
- [next → Merging Models with Fisher-Weighted Averaging](https://md.research.elamur.ai/free/en/papers/matena-2021-merging-models-fisher-weighted-averaging/)
- [↑ back to papers of 2021](https://md.research.elamur.ai/free/en/papers/2021/page/2/)

## References

39 references, 2 of them with a card of ours. ↗ leads to the work, «find» searches for one we have no id for yet.

1. David Silver, Thomas Hubert, Julian Schrittwieser, Ioannis Antonoglou, Matthew Lai, Arthur Guez, Marc Lanctot, Laurent Sifre, Dharshan Kumaran, Thore Graepel, et al. A general reinforcement learning algorithm that masters chess, shogi, and go through self-play. Science, 2018. [find](https://scholar.google.com/scholar?q=David+Silver%2C+Thomas+Hubert%2C+Julian+Schrittwieser%2C+Ioannis+Antonoglou%2C+Matthew+Lai%2C+Arthur+Guez%2C+Marc+Lanctot%2C+Laurent+Sifre%2C+Dharshan+Kumaran%2C+Thore+Graepel%2C+et+al%2E+A+general+reinforcement+learning+a)
2. Santiago Ontanón, Gabriel Synnaeve, Alberto Uriarte, Florian Richoux, David Churchill, and Mike Preuss. A survey of real-time strategy game ai research and competition in starcraft. IEEE Transactions on Computational Intelligence and AI in games, 5(4):293–311, 2013. [find](https://scholar.google.com/scholar?q=Santiago+Ontan%C3%B3n%2C+Gabriel+Synnaeve%2C+Alberto+Uriarte%2C+Florian+Richoux%2C+David+Churchill%2C+and+Mike+Preuss%2E+A+survey+of+real%2Dtime+strategy+game+ai+research+and+competition+in+starcraft%2E+IEEE+Transactions+)
3. Christopher Berner, Greg Brockman, Brooke Chan, Vicki Cheung, Przemysław D˛ebiak, Christy Dennison, David Farhi, Quirin Fischer, Shariq Hashme, Chris Hesse, et al. Dota 2 with large scale deep reinforcement learning. arXiv preprint arXiv:1912.06680, 2019. [find](https://scholar.google.com/scholar?q=Christopher+Berner%2C+Greg+Brockman%2C+Brooke+Chan%2C+Vicki+Cheung%2C+Przemys%C5%82aw+D%CB%9Bebiak%2C+Christy+Dennison%2C+David+Farhi%2C+Quirin+Fischer%2C+Shariq+Hashme%2C+Chris+Hesse%2C+et+al%2E+Dota+2+with+large+scale+deep+reinfor)
4. Xue Bin Peng, Erwin Coumans, Tingnan Zhang, Tsang-Wei Edward Lee, Jie Tan, and Sergey Levine. Learning Agile Robotic Locomotion Skills by Imitating Animals. In Robotics: Science and Systems, 07 2020. doi: 10.15607/RSS.2020.XVI.064. — [doi:10.15607/rss.2020.xvi.064](https://doi.org/10.15607/rss.2020.xvi.064) [find](https://scholar.google.com/scholar?q=Xue+Bin+Peng%2C+Erwin+Coumans%2C+Tingnan+Zhang%2C+Tsang%2DWei+Edward+Lee%2C+Jie+Tan%2C+and+Sergey+Levine%2E+Learning+Agile+Robotic+Locomotion+Skills+by+Imitating+Animals%2E+In+Robotics%3A+Science+and+Systems%2C+07+2020%2E+)
5. OpenAI, Marcin Andrychowicz, Bowen Baker, Maciek Chociej, Rafal Józefowicz, Bob McGrew, Jakub W. Pachocki, Jakub Pachocki, Arthur Petron, Matthias Plappert, Glenn Powell, Alex Ray, Jonas Schneider, Szymon Sidor, Josh Tobin, Peter Welinder, Lilian Weng, and Wojciech Zaremba. Learning dexterous in-han… [find](https://scholar.google.com/scholar?q=OpenAI%2C+Marcin+Andrychowicz%2C+Bowen+Baker%2C+Maciek+Chociej%2C+Rafal+J%C3%B3zefowicz%2C+Bob+McGrew%2C+Jakub+W%2E+Pachocki%2C+Jakub+Pachocki%2C+Arthur+Petron%2C+Matthias+Plappert%2C+Glenn+Powell%2C+Alex+Ray%2C+Jonas+Schneider%2C+Sz)
6. Emanuel Todorov, Tom Erez, and Yuval Tassa. Mujoco: A physics engine for model-based control. In 2012 IEEE/RSJ International Conference on Intelligent Robots and Systems, pages 5026–5033. IEEE, 2012. [find](https://scholar.google.com/scholar?q=Emanuel+Todorov%2C+Tom+Erez%2C+and+Yuval+Tassa%2E+Mujoco%3A+A+physics+engine+for+model%2Dbased+control%2E+In+2012+IEEE%2FRSJ+International+Conference+on+Intelligent+Robots+and+Systems%2C+pages+5026%E2%80%935033%2E+IEEE%2C+2012%2E)
7. Erwin Coumans and Yunfei Bai. Pybullet, a python module for physics simulation for games, robotics and machine learning, 2016. URL [http://pybullet.org](http://pybullet.org), 2016. [↗](http://pybullet.org)
8. Jeongseok Lee, Michael X Grey, Sehoon Ha, Tobias Kunz, Sumit Jain, Yuting Ye, Siddhartha S Srinivasa, Mike Stilman, and C Karen Liu. Dart: Dynamic animation and robotics toolkit. Journal of Open Source Software, 2018. [find](https://scholar.google.com/scholar?q=Jeongseok+Lee%2C+Michael+X+Grey%2C+Sehoon+Ha%2C+Tobias+Kunz%2C+Sumit+Jain%2C+Yuting+Ye%2C+Siddhartha+S+Srinivasa%2C+Mike+Stilman%2C+and+C+Karen+Liu%2E+Dart%3A+Dynamic+animation+and+robotics+toolkit%2E+Journal+of+Open+Sourc)
9. Russ Tedrake and the Drake Development Team. Drake: Model-based design and verification for robotics, 2019. URL [https://drake.mit.edu](https://drake.mit.edu). [↗](https://drake.mit.edu)
10. Eric Rohmer, Surya PN Singh, and Marc Freese. V-rep: A versatile and scalable robot simulation framework. In 2013 IEEE/RSJ International Conference on Intelligent Robots and Systems, pages 1321–1326. IEEE, 2013. [find](https://scholar.google.com/scholar?q=Eric+Rohmer%2C+Surya+PN+Singh%2C+and+Marc+Freese%2E+V%2Drep%3A+A+versatile+and+scalable+robot+simulation+framework%2E+In+2013+IEEE%2FRSJ+International+Conference+on+Intelligent+Robots+and+Systems%2C+pages+1321%E2%80%931326%2E+)
11. Ilge Akkaya, Marcin Andrychowicz, Maciek Chociej, Mateusz Litwin, Bob McGrew, Arthur Petron, Alex Paino, Matthias Plappert, Glenn Powell, Raphael Ribas, et al. Solving Rubik’s Cube with a Robot Hand. arXiv preprint arXiv:1910.07113, 2019. [find](https://scholar.google.com/scholar?q=Ilge+Akkaya%2C+Marcin+Andrychowicz%2C+Maciek+Chociej%2C+Mateusz+Litwin%2C+Bob+McGrew%2C+Arthur+Petron%2C+Alex+Paino%2C+Matthias+Plappert%2C+Glenn+Powell%2C+Raphael+Ribas%2C+et+al%2E+Solving+Rubik%E2%80%99s+Cube+with+a+Robot+Hand%2E+)
12. Jacky Liang, Viktor Makoviychuk, Ankur Handa, Nuttapong Chentanez, Miles Macklin, and Dieter Fox. Gpu-accelerated robotic simulation for distributed reinforcement learning. In Conference on Robot Learning. PMLR, 2018. [find](https://scholar.google.com/scholar?q=Jacky+Liang%2C+Viktor+Makoviychuk%2C+Ankur+Handa%2C+Nuttapong+Chentanez%2C+Miles+Macklin%2C+and+Dieter+Fox%2E+Gpu%2Daccelerated+robotic+simulation+for+distributed+reinforcement+learning%2E+In+Conference+on+Robot+Lear)
13. NVIDIA. Nvidia PhysX, 2020. URL [https://developer.nvidia.com/physx-sdk](https://developer.nvidia.com/physx-sdk). [↗](https://developer.nvidia.com/physx-sdk)
14. C. Daniel Freeman, Erik Frey, Anton Raichuk, Sertan Girgin, Igor Mordatch, and Olivier Bachem. Brax - A Differentiable Physics Engine for Large Scale Rigid Body Simulation, 2021. URL [http://github.com/google/brax](http://github.com/google/brax). [↗](http://github.com/google/brax)
15. Joshua Tobin, Rachel Fong, Alex Ray, Jonas Schneider, Wojciech Zaremba, and Pieter Abbeel. Domain randomization for transferring deep neural networks from simulation to the real world. CoRR, abs/1703.06907, 2017. URL http://arxiv.org/abs/1703.06907. [↗](http://arxiv.org/abs/1703.06907)
16. M. Hutter, Christian Gehring, Dominic Jud, Andreas Lauber, Dario Bellicoso, Vassilios Tsounis, Jemin Hwangbo, K. Bodie, P. Fankhauser, Michael Bloesch, Remo Diethelm, Samuel Bachmann, A. Melzer, and M. Höpflinger. Anymal - a highly mobile and dynamic quadrupedal robot. (IROS), 2016. [find](https://scholar.google.com/scholar?q=M%2E+Hutter%2C+Christian+Gehring%2C+Dominic+Jud%2C+Andreas+Lauber%2C+Dario+Bellicoso%2C+Vassilios+Tsounis%2C+Jemin+Hwangbo%2C+K%2E+Bodie%2C+P%2E+Fankhauser%2C+Michael+Bloesch%2C+Remo+Diethelm%2C+Samuel+Bachmann%2C+A%2E+Melzer%2C+and+M)
17. Xue Bin Peng, Ze Ma, Pieter Abbeel, Sergey Levine, and Angjoo Kanazawa. AMP: Adversarial Motion Priors for Stylized Physics-Based Character Control. ACM Trans. Graph., 2021. [find](https://scholar.google.com/scholar?q=Xue+Bin+Peng%2C+Ze+Ma%2C+Pieter+Abbeel%2C+Sergey+Levine%2C+and+Angjoo+Kanazawa%2E+AMP%3A+Adversarial+Motion+Priors+for+Stylized+Physics%2DBased+Character+Control%2E+ACM+Trans%2E+Graph%2E%2C+2021%2E)
18. Miles Macklin, Kier Storey, Michelle Lu, Pierre Terdiman, Nuttapong Chentanez, Stefan Jeschke, and Matthias Müller. Small steps in physics simulation. In Proceedings of the 18th Annual ACM SIGGRAPH/Eurographics Symposium on Computer Animation, SCA ’19, New York, NY, USA, 2019. Association for Comput… — [doi:10.1145/3309486.3340247](https://doi.org/10.1145/3309486.3340247) [find](https://scholar.google.com/scholar?q=Miles+Macklin%2C+Kier+Storey%2C+Michelle+Lu%2C+Pierre+Terdiman%2C+Nuttapong+Chentanez%2C+Stefan+Jeschke%2C+and+Matthias+M%C3%BCller%2E+Small+steps+in+physics+simulation%2E+In+Proceedings+of+the+18th+Annual+ACM+SIGGRAPH%2FEu)
19. John Schulman, Filip Wolski, Prafulla Dhariwal, Alec Radford, and Oleg Klimov. Proximal Policy Optimization Algorithms, 2017. [find](https://scholar.google.com/scholar?q=John+Schulman%2C+Filip+Wolski%2C+Prafulla+Dhariwal%2C+Alec+Radford%2C+and+Oleg+Klimov%2E+Proximal+Policy+Optimization+Algorithms%2C+2017%2E)
20. Denys Makoviichuk and Viktor Makoviychuk. RL Games, 2021. URL [https://github.com/Denys88/rl_games/](https://github.com/Denys88/rl_games/). [↗](https://github.com/Denys88/rl_games/)
21. Lerrel Pinto, Marcin Andrychowicz, Peter Welinder, Wojciech Zaremba, and Pieter Abbeel. Asymmetric actor critic for image-based robot learning. CoRR, 2017. URL http://arxiv.org/abs/1710.06542. [↗](http://arxiv.org/abs/1710.06542)
22. Jemin Hwangbo, Joonho Lee, Alexey Dosovitskiy, Dario Bellicoso, Vassilios Tsounis, Vladlen Koltun, and Marco Hutter. Learning Agile and Dynamic Motor Skills for Legged Robots. Science Robotics, Jan 2019. [find](https://scholar.google.com/scholar?q=Jemin+Hwangbo%2C+Joonho+Lee%2C+Alexey+Dosovitskiy%2C+Dario+Bellicoso%2C+Vassilios+Tsounis%2C+Vladlen+Koltun%2C+and+Marco+Hutter%2E+Learning+Agile+and+Dynamic+Motor+Skills+for+Legged+Robots%2E+Science+Robotics%2C+Jan+20)
23. Anonymous. Learning to walk in minutes using massively parallel deep reinforcement learning. In Submitted to 5th Annual Conference on Robot Learning, 2021. URL [https://openreview.net/forum?id=wK2fDDJ5VcF](https://openreview.net/forum?id=wK2fDDJ5VcF). under review. [↗](https://openreview.net/forum?id=wK2fDDJ5VcF)
24. Xue Bin Peng, Pieter Abbeel, Sergey Levine, and Michiel van de Panne. Deepmimic: Example-guided deep reinforcement learning of physics-based character skills. ACM Trans. Graph., 37 (4), July 2018. doi: 10.1145/3197517.3201311. — [doi:10.1145/3197517.3201311](https://doi.org/10.1145/3197517.3201311) [find](https://scholar.google.com/scholar?q=Xue+Bin+Peng%2C+Pieter+Abbeel%2C+Sergey+Levine%2C+and+Michiel+van+de+Panne%2E+Deepmimic%3A+Example%2Dguided+deep+reinforcement+learning+of+physics%2Dbased+character+skills%2E+ACM+Trans%2E+Graph%2E%2C+37+%284%29%2C+July+2018%2E+doi)
25. O. Khatib. A unified approach for motion and force control of robot manipulators: The operational space formulation. IEEE Journal on Robotics and Automation, 3(1):43–53, 1987. doi: 10.1109/JRA.1987.1087068. [find](https://scholar.google.com/scholar?q=O%2E+Khatib%2E+A+unified+approach+for+motion+and+force+control+of+robot+manipulators%3A+The+operational+space+formulation%2E+IEEE+Journal+on+Robotics+and+Automation%2C+3%281%29%3A43%E2%80%9353%2C+1987%2E+doi%3A+10%2E1109%2FJRA%2E1987%2E10)
26. Yuke Zhu, Josiah Wong, Ajay Mandlekar, and Roberto Martín-Martín. robosuite: A modular simulation framework and benchmark for robot learning, 2020. [find](https://scholar.google.com/scholar?q=Yuke+Zhu%2C+Josiah+Wong%2C+Ajay+Mandlekar%2C+and+Roberto+Mart%C3%ADn%2DMart%C3%ADn%2E+robosuite%3A+A+modular+simulation+framework+and+benchmark+for+robot+learning%2C+2020%2E)
27. Roberto Martín-Martín, Michelle A. Lee, Rachel Gardner, Silvio Savarese, Jeannette Bohg, and Animesh Garg. Variable impedance control in end-effector space: An action space for reinforcement learning in contact-rich tasks, 2019. [find](https://scholar.google.com/scholar?q=Roberto+Mart%C3%ADn%2DMart%C3%ADn%2C+Michelle+A%2E+Lee%2C+Rachel+Gardner%2C+Silvio+Savarese%2C+Jeannette+Bohg%2C+and+Animesh+Garg%2E+Variable+impedance+control+in+end%2Deffector+space%3A+An+action+space+for+reinforcement+learning+)
28. Manuel Wüthrich, Felix Widmaier, Felix Grimminger, Joel Akpo, Shruti Joshi, Vaibhav Agrawal, Bilal Hammoud, Majid Khadiv, Miroslav Bogdanovic, Vincent Berenz, Julian Viereck, Maximilien Naveau, Ludovic Righetti, Bernhard Schölkopf, and Stefan Bauer. TriFinger: An Open-Source Robot for Learning Dexte… — [arXiv:2008.03596](https://arxiv.org/abs/2008.03596) [find](https://scholar.google.com/scholar?q=Manuel+W%C3%BCthrich%2C+Felix+Widmaier%2C+Felix+Grimminger%2C+Joel+Akpo%2C+Shruti+Joshi%2C+Vaibhav+Agrawal%2C+Bilal+Hammoud%2C+Majid+Khadiv%2C+Miroslav+Bogdanovic%2C+Vincent+Berenz%2C+Julian+Viereck%2C+Maximilien+Naveau%2C+Ludovi)
29. Arthur Allshire, Mayank Mittal, Varun Lodaya, Viktor Makoviychuk, Denys Makoviichuk, Felix Widmaier, Manuel Wuthrich, Stefan Bauer, Ankur Handa, and Animesh Garg. Transferring Dexterous Manipulation from GPU Simulation to a Remote Real-World TriFinger. CoRR, 2021. [find](https://scholar.google.com/scholar?q=Arthur+Allshire%2C+Mayank+Mittal%2C+Varun+Lodaya%2C+Viktor+Makoviychuk%2C+Denys+Makoviichuk%2C+Felix+Widmaier%2C+Manuel+Wuthrich%2C+Stefan+Bauer%2C+Ankur+Handa%2C+and+Animesh+Garg%2E+Transferring+Dexterous+Manipulation+f)
30. Niklas Funk, Charles B. Schaff, Rishabh Madan, Takuma Yoneda, Julen Urain De Jesus, Joe Watson, Ethan K. Gordon, Felix Widmaier, Stefan Bauer, Siddhartha S. Srinivasa, Tapomayukh Bhattacharjee, Matthew R. Walter, and Jan Peters. Benchmarking structured policies and policy optimization for real-world… — [arXiv:2105.02087](https://arxiv.org/abs/2105.02087) [find](https://scholar.google.com/scholar?q=Niklas+Funk%2C+Charles+B%2E+Schaff%2C+Rishabh+Madan%2C+Takuma+Yoneda%2C+Julen+Urain+De+Jesus%2C+Joe+Watson%2C+Ethan+K%2E+Gordon%2C+Felix+Widmaier%2C+Stefan+Bauer%2C+Siddhartha+S%2E+Srinivasa%2C+Tapomayukh+Bhattacharjee%2C+Matthe)
31. J. D. Hunter. Matplotlib: A 2d graphics environment. Computing in Science & Engineering, pages 90–95, 2007. [find](https://scholar.google.com/scholar?q=J%2E+D%2E+Hunter%2E+Matplotlib%3A+A+2d+graphics+environment%2E+Computing+in+Science+%26+Engineering%2C+pages+90%E2%80%9395%2C+2007%2E)
32. Guido Van Rossum and Fred L. Drake. Python 3 Reference Manual. CreateSpace, Scotts Valley, CA, 2009. [find](https://scholar.google.com/scholar?q=Guido+Van+Rossum+and+Fred+L%2E+Drake%2E+Python+3+Reference+Manual%2E+CreateSpace%2C+Scotts+Valley%2C+CA%2C+2009%2E)
33. Charles R. Harris, K. Jarrod Millman, Stéfan J. van der Walt, Ralf Gommers, Pauli Virtanen, David Cournapeau, Eric Wieser, Julian Taylor, Sebastian Berg, Nathaniel J. Smith, Robert Kern, Matti Picus, Stephan Hoyer, Marten H. van Kerkwijk, Matthew Brett, Allan Haldane, Jaime Fernández del Río, Mark W… — [Array Programming with NumPy](https://md.research.elamur.ai/free/en/papers/harris-2020-array-programming-numpy/)
34. Adam Paszke, Sam Gross, Francisco Massa, Adam Lerer, James Bradbury, Gregory Chanan, Trevor Killeen, Zeming Lin, Natalia Gimelshein, Luca Antiga, Alban Desmaison, Andreas Kopf, Edward Yang, Zachary DeVito, Martin Raison, Alykhan Tejani, Sasank Chilamkurthy, Benoit Steiner, Lu Fang, Junjie Bai, and S… — [PyTorch: An Imperative Style, High-Performance Deep Learning Library](https://md.research.elamur.ai/free/en/papers/paszke-2019-pytorch-imperative-style-high-performance/)
35. Martín Abadi, Ashish Agarwal, Paul Barham, Eugene Brevdo, Zhifeng Chen, Craig Citro, Greg S. Corrado, Andy Davis, Jeffrey Dean, Matthieu Devin, Sanjay Ghemawat, Ian Goodfellow, Andrew Harp, Geoffrey Irving, Michael Isard, Yangqing Jia, Rafal Jozefowicz, Lukasz Kaiser, Manjunath Kudlur, Josh Levenber… [find](https://scholar.google.com/scholar?q=Mart%C3%ADn+Abadi%2C+Ashish+Agarwal%2C+Paul+Barham%2C+Eugene+Brevdo%2C+Zhifeng+Chen%2C+Craig+Citro%2C+Greg+S%2E+Corrado%2C+Andy+Davis%2C+Jeffrey+Dean%2C+Matthieu+Devin%2C+Sanjay+Ghemawat%2C+Ian+Goodfellow%2C+Andrew+Harp%2C+Geoffrey+I)
36. Sebastian Penhouet. TensorBoard Aggregator, February 2021. URL [https://github.com/](https://github.com/) Spenhouet/tensorboard-aggregator. [↗](https://github.com/)
37. John D. Garrett and Hsin-Hsiang Peng. garrettj403/SciencePlots, February 2021. URL http://doi.org/10.5281/zenodo.4106649. — [doi:10.5281/zenodo.4106649](https://doi.org/10.5281/zenodo.4106649) [↗](http://doi.org/10.5281/zenodo.4106649)
38. Overleaf, 2012. URL [https://www.overleaf.com/](https://www.overleaf.com/). [↗](https://www.overleaf.com/)
39. Ossama Ahmed, Frederik Träuble, Anirudh Goyal, Alexander Neitz, Manuel Wüthrich, Yoshua Bengio, Bernhard Schölkopf, and Stefan Bauer. CausalWorld: A Robotic Manipulation Benchmark for Causal Structure and Transfer Learning. CoRR, abs/2010.04296, 2020. URL https://arxiv.org/abs/2010.04296. — [arXiv:2010.04296](https://arxiv.org/abs/2010.04296) [↗](https://arxiv.org/abs/2010.04296)

*Updated 2 November 2026*