I’m interested in scaling up reinforcement learning experiments in Isaac Sim. I’m wondering if it’s possible to spawn multiple gym environments both locally and over multiple AWS EC2 instances, and have each environment interact with a central agent that is the learner?
Related topics
Topic | Replies | Views | Activity | |
---|---|---|---|---|
Can Isaac Gym support multi-agent reinforcement learning? | 12 | 2593 | December 19, 2023 | |
Running Isaac Gym on AWS for Soft Body Simulation | 0 | 232 | April 5, 2024 | |
Isaac Sim and Gym scalability | 3 | 898 | October 10, 2021 | |
Is it possible to do Reinforcement learning research with isacc sim? | 1 | 402 | December 1, 2021 | |
Parallelized simulation | 0 | 211 | December 12, 2023 | |
Run Isaac gym on multiple machines' GPUs in parallel | 3 | 824 | June 7, 2022 | |
Is it possible to make multiple agent to interact with each other for a swarm robotics application? | 3 | 1328 | July 14, 2021 | |
Isaac sim simulating subsets of the environments | 1 | 23 | September 25, 2024 | |
How scalable is Isaac? | 1 | 67 | October 4, 2024 | |
Isaac Gym for Multi-Robot Cooperation/Exploration/Navigation in an Environment | 4 | 723 | December 25, 2023 |