add simulation benchmark

[LBG21]

No description provided.

[ethan-iai]

We are excited to announce that the simulation benchmark has been successfully integrated into ManiSkill (special thanks to the developers of ManiSkill, especially @StoneT2000) for the Robotics Diffusion Transformer (RDT)! 🎉

This PR directly addresses the following issues:

• How to run inference of the model with a single image and no proprioception data? #5: Deployment of RDT on the SimplerEnv.
• Nice! How about experimental results in some simulation-based scenes? #10 & Do you have any simulation to recommend? #32: Requests for a simulation-based benchmark for RDT.

This new benchmark significantly improves the evaluation and performance testing of RDT in simulated environments, ensuring more comprehensive results.

We encourage you to check out these PR for more details and updates. As always, your feedback is invaluable in refining the benchmark and optimizing the model’s performance in simulation environments.

Thank you for your continued support! 😊

[XieJunr]

Great! So is it possible to run inference of the model in the simulation with a single image and no proprioception data？

我们很高兴地宣布，仿真基准测试已成功集成到机器人扩散转换器 （RDT） 的 ManiSkill（特别感谢 ManiSkill 的开发人员）中！🎉

此 PR 直接解决以下问题：

• 如何使用单个图像而没有本体感觉数据运行模型的推理？ #5：在 SimplerEnv 上部署 RDT。
• 好！在一些基于模拟的场景中的实验结果怎么样？ #10 & 您有什么模拟推荐吗？ #32：请求基于模拟的 RDT 基准测试。

这个新的基准测试显著改善了模拟环境中 RDT 的评估和性能测试，确保了更全面的结果。

我们鼓励你查看这些 PR 以获取更多详细信息和更新。与往常一样，您的反馈对于优化基准测试和优化模型在仿真环境中的性能非常宝贵。

感谢您一直以来的支持！😊

[csuastt]

Yes, it should be.