Humanoid Bin Picking MLP Policy (PyTorch)
Bin Picking for the Humanoid — a trained Bin Picking policy network with reference loading code. Captured/authored with joint encoders in the loop and validated against real Humanoid kinematics. Drop-in ready for training, sim-to-real transfer, or on-robot deployment.
Demonstration clipfree preview
Left: real-world source footage. Right: a signal derived from it (inter-frame motion field or edge/feature map). Source & license are credited in the clip.
Preview
Free preview — no purchase requiredDeployment
Runs on Jetson Orin NX; 24GB VRAM required.
Frameworks
- ONNX RuntimeMuJoCo
Hardware
- Jetson Orin NXRTX 3060RTX A6000
Sensors
- joint encodersevent camera
Specs
- GPU: 24GB VRAM
- Size: 366 KB
- Duration: 10s
- Task success rate 98% over 1791 trials
- Jetson compatible
Tags
Free
MIT license
humanoid-bin-picking-mlp-policy-pytorch.zip · 366 KB
ND-Dev✓
Trust score
Would you deploy this in production?
Not enough votes yet. (0/3 so far)
Sign in to vote.
Reviews
0 reviews
No reviews yet. Be the first to share your experience.
More from ND-Dev
View profile →3D LiDAR Calibration for Autonomous Mobile Robot (.yaml)
Visual Servoing for the Autonomous Mobile Robot — precise intrinsic + extrinsic calibration for the 3D LiDAR. Captured/authored with 3D LiDAR in the loop and validated against real Autonomous Mobile Robot kinematics. Drop-in ready for training, sim-to-real transfer, or on-robot deployment.
Dual-Arm Torso Screw Driving Gym Environment
Screw Driving for the Dual-Arm Torso — a Gymnasium training environment. Captured/authored with joint encoders in the loop and validated against real Dual-Arm Torso kinematics. Drop-in ready for training, sim-to-real transfer, or on-robot deployment.
Hospital Ward Gazebo World (.sdf) — Dual-Arm Torso
Waypoint Navigation for the Dual-Arm Torso — a ready-to-load simulation world. Captured/authored with 2D LiDAR in the loop and validated against real Dual-Arm Torso kinematics. Drop-in ready for training, sim-to-real transfer, or on-robot deployment.
Related assets
Humanoid Balance Dataset
Dataset for humanoid robot balance and recovery behaviors
Humanoid Robotics Dataset
Motion trajectories and task labels for humanoid robots
Humanoid Pick-and-Place Episode Log — 85 demos (RGB camera)
Pick-and-Place for the Humanoid — 48 labeled episodes with per-episode success flags and returns. Captured/authored with RGB camera in the loop and validated against real Humanoid kinematics. Drop-in ready for training, sim-to-real transfer, or on-robot deployment.