Firas Al-Hafez
@firasalhafez.bsky.social
Ph.D. student in machine learning and robotics.
firasalhafez.com
firasalhafez.com
Work conducted @ias-tudarmstadt.bsky.social headed by @jan-peters.bsky.social @cs-tudarmstadt.bsky.social @tuda.bsky.social
Many thanks to all #LocoMuJoCo Contributors!
Many thanks to all #LocoMuJoCo Contributors!
April 18, 2025 at 10:47 PM
Work conducted @ias-tudarmstadt.bsky.social headed by @jan-peters.bsky.social @cs-tudarmstadt.bsky.social @tuda.bsky.social
Many thanks to all #LocoMuJoCo Contributors!
Many thanks to all #LocoMuJoCo Contributors!
🦿 LocoMuJoCo is the first benchmark bridging biomechanics and humanoid robotics!
Train biomechanical models in parallel on GPU using our JAX-based pipeline.
We provide models of varying complexity to accelerate biomechanics research.
Train biomechanical models in parallel on GPU using our JAX-based pipeline.
We provide models of varying complexity to accelerate biomechanics research.
April 18, 2025 at 10:47 PM
🦿 LocoMuJoCo is the first benchmark bridging biomechanics and humanoid robotics!
Train biomechanical models in parallel on GPU using our JAX-based pipeline.
We provide models of varying complexity to accelerate biomechanics research.
Train biomechanical models in parallel on GPU using our JAX-based pipeline.
We provide models of varying complexity to accelerate biomechanics research.
🤖 We provide a robot2robot retargeting pipeline!
Retarget any trajectory to any humanoid — regardless of DoF or size.
Add a new model, and get instant access to 22K+ trajectories for training!
Retarget any trajectory to any humanoid — regardless of DoF or size.
Add a new model, and get instant access to 22K+ trajectories for training!
April 18, 2025 at 10:47 PM
🤖 We provide a robot2robot retargeting pipeline!
Retarget any trajectory to any humanoid — regardless of DoF or size.
Add a new model, and get instant access to 22K+ trajectories for training!
Retarget any trajectory to any humanoid — regardless of DoF or size.
Add a new model, and get instant access to 22K+ trajectories for training!
⚡️ Lightning-fast training by JIT-compiling the environment and the algorithm into a single function!
Train the policy below in under 20 minutes on an RTX 4090.
Train the policy below in under 20 minutes on an RTX 4090.
April 18, 2025 at 10:47 PM
⚡️ Lightning-fast training by JIT-compiling the environment and the algorithm into a single function!
Train the policy below in under 20 minutes on an RTX 4090.
Train the policy below in under 20 minutes on an RTX 4090.