We will train two skills for the manipulator
- Peg Insertion
- Gear Meshing
- Go to your IsaacLab directory
source/isaaclab_tasks/isaaclab_tasks/manager_based - Clone the repository
https://github.com/bugartlan/HCR_Gear_Train_Assembly.git
- UR3e
- Robotiq Hand-E
- Train on HCRL servers: Same command as how you run SO_101, except replacing the task name with
./ray.sh job --task Isaac-Assembly-PegInsert-Chamfer-v0- Run training without visualization (headless): Use
--num_env 1for faster startup for quick syntax checks.
python scripts/reinforcement_learning/rsl_rl/train.py --task Isaac-Assembly-PegInsert-v0 --headless- Run training with visualization for visual debugging.
python scripts/reinforcement_learning/rsl_rl/train.py --task Isaac-Assembly-PegInsert-v0 --num_env 1- Export and run the most recent trained policy:
python scripts/reinforcement_learning/rsl_rl/play.py --task Isaac-Assembly-PegInsert-v0 --num_env 1