AgiBot Lingxi X1 · published recipe
x1_dh_stand training recipe (standing/walking, 12 leg actions)
Record
- Robot
- AgiBot Lingxi X1 agibot/x1
- Framework
- Isaac Gym Preview 4 + rsl_rl-style PPO (legged_gym / humanoid-gym lineage) · Isaac Gym
- Policy license
- not stated by the package · robot software repos: MulanPSL-2.0
- File
- humanoid/scripts/train.py
- Note
- No trained checkpoint in the repo; export_onnx_dh.py produces the ONNX that the infer stack loads.
Not in the tab yet
To replay this policy in the browser we need an ONNX export of a trained checkpoint, the observation and action contract copied from the authors' own sim2sim or deploy script, and a MuJoCo model of the robot that the contract matches. Then it must walk 60 s in our runtime without a fall. Until then the source link is the way in.
Other AgiBot Lingxi X1 policies · 2
- walk, legs only (upper body held still), joystick velocity commandsONNX Runtime in AimRT rl_control_module, 1 kHz control; trained with agibot_x1_train (Isaac Gym, humanoid-gym style PPO)
- walk with shoulder pitch following the gait (walk_leg_arm mode)ONNX Runtime in AimRT rl_control_module