MuJoCo Warp #2511
Replies: 11 comments
-
Can't wait! |
Beta Was this translation helpful? Give feedback.
-
🚀🚀🚀🚀 |
Beta Was this translation helpful? Give feedback.
-
Lets go! |
Beta Was this translation helpful? Give feedback.
-
Awesome! |
Beta Was this translation helpful? Give feedback.
-
Thanks for this work, struggle on mjx backward for large scene and object for a long time!!! |
Beta Was this translation helpful? Give feedback.
-
🚀🚀🚀 |
Beta Was this translation helpful? Give feedback.
-
🤩 |
Beta Was this translation helpful? Give feedback.
-
🚀🚀🚀 |
Beta Was this translation helpful? Give feedback.
-
Will MJWarp produce faster finite element-based simulations (e.g. using MuJoCo's elasticity model)? |
Beta Was this translation helpful? Give feedback.
-
Nice!! Give me that link 🤩 |
Beta Was this translation helpful? Give feedback.
-
Can't wait! |
Beta Was this translation helpful? Give feedback.
Uh oh!
There was an error while loading. Please reload this page.
Uh oh!
There was an error while loading. Please reload this page.
-
Google DeepMind and NVIDIA have teamed up on a new open source implementation of MuJoCo called MuJoCo Warp (MJWarp)!
MJWarp is a GPU-optimized version of MuJoCo, designed for NVIDIA hardware. MJWarp uses NVIDIA Warp to circumvent many of the sharp bits in MuJoCo MJX - MuJoCo simulations of more complex scenes (such as manipulation tasks) run >100x faster in MJWarp compared to MJX today.
For those attending NVIDIA GTC in San Jose, we will share some details of MuJoCo Warp at the following talk: Physical AI for Humanoids: How Google Robotics Uses Simulation to Accelerate Humanoid Robotics Training
MuJoCo Warp is now available at https://github.com/google-deepmind/mujoco_warp
The code is in an early Alpha version with many features still missing. Once MJWarp exits Alpha, it will be integrated into MJX - anyone who uses MJX today will instantly benefit.
Look forward to seeing you at GTC!
Beta Was this translation helpful? Give feedback.
All reactions