Example gallery#
Below is a gallery of examples
1. Data collection and processing#
The following examples show how to collect multimodal demonstration data.
2. Reinforcement learning class#
The following are examples of reinforcement learning methods for robot learning.
'''Training'''
python examples/learning_rl/IsaacGym_RofuncRL/example_Ant_RofuncRL.py --agent=[ppo|a2c|td3|sac]
'''Inference with pre-trained model in model zoo'''
python examples/learning_rl/IsaacGym_RofuncRL/example_Ant_RofuncRL.py --agent=ppo --inference
'''Training'''
python examples/learning_rl/IsaacGym_RofuncRL/example_CURICabinet_RofuncRL.py --agent=ppo
'''Inference with pre-trained model in model zoo'''
python examples/learning_rl/IsaacGym_RofuncRL/example_CURICabinet_RofuncRL.py --agent=ppo --inference
# Available objects: Hammer, Spatula, Large_Clamp, Mug, Power_Drill, Knife, Scissors, Large_Marker, Phillips_Screw_Driver
'''Training'''
python examples/learning_rl/IsaacGym_RofuncRL/example_DexterousHands_RofuncRL.py --task=CURIQbSoftHandSynergyGrasp --agent=ppo --objects=Hammer
'''Inference with pre-trained model in model zoo'''
python examples/learning_rl/IsaacGym_RofuncRL/example_DexterousHands_RofuncRL.py --task=CURIQbSoftHandSynergyGrasp --agent=ppo --inference --objects=Hammer
'''Training'''
python examples/learning_rl/IsaacGym_RofuncRL/example_FrankaCabinet_RofuncRL.py --agent=ppo
'''Inference with pre-trained model in model zoo'''
python examples/learning_rl/IsaacGym_RofuncRL/example_FrankaCabinet_RofuncRL.py --agent=ppo --inference
'''Training'''
python examples/learning_rl/IsaacGym_RofuncRL/example_FrankaCubeStack_RofuncRL.py --agent=ppo
'''Training'''
python examples/learning_rl/IsaacGym_RofuncRL/example_Humanoid_RofuncRL.py --agent=ppo
'''Inference with pre-trained model in model zoo'''
python examples/learning_rl/IsaacGym_RofuncRL/example_Humanoid_RofuncRL.py --agent=ppo --inference
'''Training'''
# Backflip
python examples/learning_rl/IsaacGym_RofuncRL/example_HumanoidAMP_RofuncRL.py --task=HumanoidAMP_backflip --agent=amp
# Walk
python examples/learning_rl/IsaacGym_RofuncRL/example_HumanoidAMP_RofuncRL.py --task=HumanoidAMP_walk --agent=amp
# Run
python examples/learning_rl/IsaacGym_RofuncRL/example_HumanoidAMP_RofuncRL.py --task=HumanoidAMP_run --agent=amp
# Dance
python examples/learning_rl/IsaacGym_RofuncRL/example_HumanoidAMP_RofuncRL.py --task=HumanoidAMP_dance --agent=amp
# Hop
python examples/learning_rl/IsaacGym_RofuncRL/example_HumanoidAMP_RofuncRL.py --task=HumanoidAMP_hop --agent=amp
'''Inference with pre-trained model in model zoo'''
# Backflip
python examples/learning_rl/IsaacGym_RofuncRL/example_HumanoidAMP_RofuncRL.py --task=HumanoidAMP_backflip --agent=amp --inference
# Walk
python examples/learning_rl/IsaacGym_RofuncRL/example_HumanoidAMP_RofuncRL.py --task=HumanoidAMP_walk --agent=amp --inference
# Run
python examples/learning_rl/IsaacGym_RofuncRL/example_HumanoidAMP_RofuncRL.py --task=HumanoidAMP_run --agent=amp --inference
# Dance
python examples/learning_rl/IsaacGym_RofuncRL/example_HumanoidAMP_RofuncRL.py --task=HumanoidAMP_dance --agent=amp --inference
# Hop
python examples/learning_rl/IsaacGym_RofuncRL/example_HumanoidAMP_RofuncRL.py --task=HumanoidAMP_hop --agent=amp --inference
'''Training'''
# Getup
python examples/learning_rl/IsaacGym_RofuncRL/example_HumanoidASE_RofuncRL.py --task=HumanoidASEGetupSwordShield --agent=ase
# Getup with perturbation
python examples/learning_rl/IsaacGym_RofuncRL/example_HumanoidASE_RofuncRL.py --task=HumanoidASEPerturbSwordShield --agent=ase
# Heading
python examples/learning_rl/IsaacGym_RofuncRL/example_HumanoidASE_RofuncRL.py --task=HumanoidASEHeadingSwordShield --agent=ase
# Reach
python examples/learning_rl/IsaacGym_RofuncRL/example_HumanoidASE_RofuncRL.py --task=HumanoidASEReachSwordShield --agent=ase
# Location
python examples/learning_rl/IsaacGym_RofuncRL/example_HumanoidASE_RofuncRL.py --task=HumanoidASELocationSwordShield --agent=ase
# Strike
python examples/learning_rl/IsaacGym_RofuncRL/example_HumanoidASE_RofuncRL.py --task=HumanoidASEStrikeSwordShield --agent=ase
'''Inference with pre-trained model in model zoo'''
# Getup
python examples/learning_rl/IsaacGym_RofuncRL/example_HumanoidASE_RofuncRL.py --task=HumanoidASEGetupSwordShield --agent=ase --inference
# Getup with perturbation
python examples/learning_rl/IsaacGym_RofuncRL/example_HumanoidASE_RofuncRL.py --task=HumanoidASEPerturbSwordShield --agent=ase --inference
# Heading
python examples/learning_rl/IsaacGym_RofuncRL/example_HumanoidASE_RofuncRL.py --task=HumanoidASEHeadingSwordShield --agent=ase --inference
# Reach
python examples/learning_rl/IsaacGym_RofuncRL/example_HumanoidASE_RofuncRL.py --task=HumanoidASEReachSwordShield --agent=ase --inference
# Location
python examples/learning_rl/IsaacGym_RofuncRL/example_HumanoidASE_RofuncRL.py --task=HumanoidASELocationSwordShield --agent=ase --inference
# Strike
python examples/learning_rl/IsaacGym_RofuncRL/example_HumanoidASE_RofuncRL.py --task=HumanoidASEStrikeSwordShield --agent=ase --inference
'''Training'''
# Available tasks: BiShadowHandOver, BiShadowHandBlockStack, BiShadowHandBottleCap, BiShadowHandCatchAbreast,
# BiShadowHandCatchOver2Underarm, BiShadowHandCatchUnderarm, BiShadowHandDoorOpenInward,
# BiShadowHandDoorOpenOutward, BiShadowHandDoorCloseInward, BiShadowHandDoorCloseOutward,
# BiShadowHandGraspAndPlace, BiShadowHandLiftUnderarm, BiShadowHandPen, BiShadowHandPointCloud,
# BiShadowHandPushBlock, BiShadowHandReOrientation, BiShadowHandScissors, BiShadowHandSwingCup,
# BiShadowHandSwitch, BiShadowHandTwoCatchUnderarm
python examples/learning_rl/IsaacGym_RofuncRL/example_DexterousHands_RofuncRL.py --task=BiShadowHandOver --agent=ppo
'''Inference with pre-trained model in model zoo'''
python examples/learning_rl/IsaacGym_RofuncRL/example_DexterousHands_RofuncRL.py --task=BiShadowHandOver --agent=ppo --inference
Tasks |
Animation |
Performance |
|
---|---|---|---|
✅ |
|||
✅ |
|||
✅ |
|||
✅ |
|||
✅ |
|||
✅ |
|||
✅ |
|||
✅ |
|||
✅ |
|||
✅ |
|||
✅ |
|||
✅ |
|||
✅ |
|||
✅ |
|||
✅ |
|||
✅ |
|||
✅ |
|||
✅ |
|||
✅ |
|||
✅ |
|||
✅ |
|||
✅ |
|||
✅ |
|||
✅ |
|||
✅ |
|||
✅ |
|||
✅ |
|||
✅ |
|||
✅ |
|||
✅ |
|||
✅ |
|||
✅ |
|||
✅ |
|||
✅ |
|||
✅ |
'''Training'''
python examples/learning_rl/OmniIsaacGym_RofuncRL/example_AllegroHandOmni_RofuncRL.py --agent=ppo
'''Training'''
python examples/learning_rl/OmniIsaacGym_RofuncRL/example_AntOmni_RofuncRL.py --agent=ppo
'''Training'''
python examples/learning_rl/OmniIsaacGym_RofuncRL/example_AnymalOmni_RofuncRL.py --agent=ppo
'''Training'''
python examples/learning_rl/OmniIsaacGym_RofuncRL/example_AnymalTerrainOmni_RofuncRL.py --agent=ppo
'''Training'''
python examples/learning_rl/OmniIsaacGym_RofuncRL/example_BallBalanceOmni_RofuncRL.py --agent=ppo
'''Training'''
python examples/learning_rl/OmniIsaacGym_RofuncRL/example_CartpoleOmni_RofuncRL.py --agent=ppo
'''Training'''
python examples/learning_rl/OmniIsaacGym_RofuncRL/example_CrazyflieOmni_RofuncRL.py --agent=ppo
'''Training'''
python examples/learning_rl/OmniIsaacGym_RofuncRL/example_FactoryNutBoltPickOmni_RofuncRL.py --agent=ppo
'''Training'''
python examples/learning_rl/OmniIsaacGym_RofuncRL/example_FrankaCabinetOmni_RofuncRL.py --agent=ppo
'''Training'''
python examples/learning_rl/OmniIsaacGym_RofuncRL/example_HumanoidOmni_RofuncRL.py --agent=ppo
'''Training'''
python examples/learning_rl/OmniIsaacGym_RofuncRL/example_IngenuityOmni_RofuncRL.py --agent=ppo
'''Training'''
python examples/learning_rl/OmniIsaacGym_RofuncRL/example_QuadcopterOmni_RofuncRL.py --agent=ppo
'''Training'''
python examples/learning_rl/OmniIsaacGym_RofuncRL/example_ShadowHandOmni_RofuncRL.py --agent=ppo
'''Training'''
python examples/learning_rl/OpenAIGym_RofuncRL/example_GymTasks_RofuncRL.py --task=Gym_Pendulum-v1 --agent=[ppo|a2c|td3|sac]
'''Training'''
python examples/learning_rl/OpenAIGym_RofuncRL/example_GymTasks_RofuncRL.py --task=Gym_CartPole-v1 --agent=[ppo|a2c|td3|sac]
'''Training'''
python examples/learning_rl/OpenAIGym_RofuncRL/example_GymTasks_RofuncRL.py --task=Gym_Acrobot-v1 --agent=[ppo|a2c|td3|sac]
'''Training'''
# Hopper
python examples/learning_rl/D4RL_Rofunc/example_D4RL_RofuncRL.py --task=Hopper --agent=dtrans
# Walker2d
python examples/learning_rl/D4RL_Rofunc/example_D4RL_RofuncRL.py --task=Walker2d --agent=dtrans
# HalfCheetah
python examples/learning_rl/D4RL_Rofunc/example_D4RL_RofuncRL.py --task=HalfCheetah --agent=dtrans
# Reacher2d
python examples/learning_rl/D4RL_Rofunc/example_D4RL_RofuncRL.py --task=Reacher2d --agent=dtrans
3. Machine learning class#
The following are examples of machine learning methods for robot learning.
TP-GMMBi with Relative Parameterization in Control
TP-GMMBi with Relative Parameterization in Representation
4. Planning and control methods#
The following are examples of planning and control methods that can be used to.
LQT with control primitives and DMP
iLQR control primitive version
5. Robolab#
The following examples show the use of the Robolab API.
CURI FK transformation verification with optitrack
6. Visualab#
The following examples show the use of the Visualab API.
Image segmentation using EfficientSAM with prompt
Image segmentation using SAM with prompt
Part-level segmentation using SAM and VLPart with prompt
7. Simulator#
The following examples show how to use the simulator.
Apply Forces and Torques On CURI
Construct custom human model from Xsens data
LLM control a robot with pre-defined low-level motion API
Object Simulator and Camera Tracking
Tracking the trajectory by Franka
Tracking the trajectory by Gluon
Tracking the trajectory with interference by CURI
Tracking the trajectory with multiple rigid bodies by CURI
Tracking the trajectory with multiple rigid bodies by Walker
Tracking the trajectory with multiple rigid bodies by Walker