AuroraMiddleware/bullet3
1
0
Fork 0
mirror of https://github.com/bulletphysics/bullet3 synced 2024-12-15 14:10:11 +00:00
Code Issues Releases Wiki Activity
851ca5bfb3
bullet3/examples/pybullet/gym/pybullet_envs/__init__.py

166 lines
3.7 KiB
Python
Raw Normal View History

add gym examples
2017-03-10 19:22:38 +00:00
from gym.envs.registration import registry, register, make, spec
# ------------bullet-------------
Improve PyBullet ports of Roboschool envs: fix reset (it kept adding stadium objects, causing slowdown), now reset uses saveState/restoreState and reset becomes a few orders of magnitude faster. Use python -m pybullet_envs.examples.testEnv --env AntBulletEnv-v0 --render=1 --steps 1000 --resetbenchmark=1 Added environments: HumanoidFlagrunBulletEnv-v0, HumanoidFlagrunHarderBulletEnv-v0, StrikerBulletEnv-v0, ThrowerBulletEnv-v0, PusherBulletEnv-v0, ReacherBulletEnv-v0, CartPoleBulletEnv-v0 and register them to OpenAI Gym. Allow numpy/humanoid_running.py to use abtch or non-batch update (setJointMotorControl2/setJointMotorControlArray)
2018-01-15 20:48:32 +00:00
register(
id='CartPoleBulletEnv-v0',
entry_point='pybullet_envs.bullet:CartPoleBulletEnv',
timestep_limit=1000,
reward_threshold=950.0,
)
add gym examples
2017-03-10 19:22:38 +00:00
register(
id='MinitaurBulletEnv-v0',
refactor pybullet/gym to allow instantiating environments directly from a pybullet install: work-in-progress (need to add missing data files, fix paths etc) example: pip install pybullet pip install gym python import gym import pybullet import pybullet_envs env = gym.make("HumanoidBulletEnv-v0")
2017-08-22 07:42:02 +00:00
entry_point='pybullet_envs.bullet:MinitaurBulletEnv',
add gym examples
2017-03-10 19:22:38 +00:00
timestep_limit=1000,
reward_threshold=5.0,
)
add RacecarGymEnv as a gym experimentation environment
2017-06-09 02:45:48 +00:00
add minitaur duck environment see also https://www.youtube.com/watch?v=SLwUCMZdKLo and evolution strategies (hardmaru): https://twitter.com/hardmaru/status/925074585984237568
2017-10-31 23:05:52 +00:00
register(
id='MinitaurBulletDuckEnv-v0',
entry_point='pybullet_envs.bullet:MinitaurBulletDuckEnv',
timestep_limit=1000,
reward_threshold=5.0,
)
add RacecarGymEnv as a gym experimentation environment
2017-06-09 02:45:48 +00:00
register(
id='RacecarBulletEnv-v0',
fixes in racecarGymEnv: implement 'render' rgb image, fix in naming, fix in observation bounds.
2017-08-25 05:01:45 +00:00
entry_point='pybullet_envs.bullet:RacecarGymEnv',
add RacecarGymEnv as a gym experimentation environment
2017-06-09 02:45:48 +00:00
timestep_limit=1000,
reward_threshold=5.0,
)
prepare to train racecar using ZED camera pixels (CNN+DQN)
2017-06-13 23:04:50 +00:00
register(
Improve PyBullet ports of Roboschool envs: fix reset (it kept adding stadium objects, causing slowdown), now reset uses saveState/restoreState and reset becomes a few orders of magnitude faster. Use python -m pybullet_envs.examples.testEnv --env AntBulletEnv-v0 --render=1 --steps 1000 --resetbenchmark=1 Added environments: HumanoidFlagrunBulletEnv-v0, HumanoidFlagrunHarderBulletEnv-v0, StrikerBulletEnv-v0, ThrowerBulletEnv-v0, PusherBulletEnv-v0, ReacherBulletEnv-v0, CartPoleBulletEnv-v0 and register them to OpenAI Gym. Allow numpy/humanoid_running.py to use abtch or non-batch update (setJointMotorControl2/setJointMotorControlArray)
2018-01-15 20:48:32 +00:00
id='RacecarZedBulletEnv-v0',
entry_point='pybullet_envs.bullet:RacecarZEDGymEnv',
timestep_limit=1000,
reward_threshold=5.0,
prepare to train racecar using ZED camera pixels (CNN+DQN)
2017-06-13 23:04:50 +00:00
)
add humanoid and kuka gym environments (experimental)
2017-06-14 07:54:41 +00:00
Rename old humanoid gym to simple humanoid gym.
2017-08-16 22:06:01 +00:00
register(
Improve PyBullet ports of Roboschool envs: fix reset (it kept adding stadium objects, causing slowdown), now reset uses saveState/restoreState and reset becomes a few orders of magnitude faster. Use python -m pybullet_envs.examples.testEnv --env AntBulletEnv-v0 --render=1 --steps 1000 --resetbenchmark=1 Added environments: HumanoidFlagrunBulletEnv-v0, HumanoidFlagrunHarderBulletEnv-v0, StrikerBulletEnv-v0, ThrowerBulletEnv-v0, PusherBulletEnv-v0, ReacherBulletEnv-v0, CartPoleBulletEnv-v0 and register them to OpenAI Gym. Allow numpy/humanoid_running.py to use abtch or non-batch update (setJointMotorControl2/setJointMotorControlArray)
2018-01-15 20:48:32 +00:00
id='SimpleHumanoidBulletEnv-v0',
entry_point='pybullet_envs.bullet:SimpleHumanoidGymEnv',
timestep_limit=1000,
reward_threshold=5.0,
Rename old humanoid gym to simple humanoid gym.
2017-08-16 22:06:01 +00:00
)
add humanoid and kuka gym environments (experimental)
2017-06-14 07:54:41 +00:00
register(
Improve PyBullet ports of Roboschool envs: fix reset (it kept adding stadium objects, causing slowdown), now reset uses saveState/restoreState and reset becomes a few orders of magnitude faster. Use python -m pybullet_envs.examples.testEnv --env AntBulletEnv-v0 --render=1 --steps 1000 --resetbenchmark=1 Added environments: HumanoidFlagrunBulletEnv-v0, HumanoidFlagrunHarderBulletEnv-v0, StrikerBulletEnv-v0, ThrowerBulletEnv-v0, PusherBulletEnv-v0, ReacherBulletEnv-v0, CartPoleBulletEnv-v0 and register them to OpenAI Gym. Allow numpy/humanoid_running.py to use abtch or non-batch update (setJointMotorControl2/setJointMotorControlArray)
2018-01-15 20:48:32 +00:00
id='KukaBulletEnv-v0',
entry_point='pybullet_envs.bullet:KukaGymEnv',
timestep_limit=1000,
reward_threshold=5.0,
add kukaCamGymEnv.py with camera observations (preliminary) show camera position in example browser disable per-vertex and per-fragment profile timings
2017-06-21 16:33:46 +00:00
)
register(
Improve PyBullet ports of Roboschool envs: fix reset (it kept adding stadium objects, causing slowdown), now reset uses saveState/restoreState and reset becomes a few orders of magnitude faster. Use python -m pybullet_envs.examples.testEnv --env AntBulletEnv-v0 --render=1 --steps 1000 --resetbenchmark=1 Added environments: HumanoidFlagrunBulletEnv-v0, HumanoidFlagrunHarderBulletEnv-v0, StrikerBulletEnv-v0, ThrowerBulletEnv-v0, PusherBulletEnv-v0, ReacherBulletEnv-v0, CartPoleBulletEnv-v0 and register them to OpenAI Gym. Allow numpy/humanoid_running.py to use abtch or non-batch update (setJointMotorControl2/setJointMotorControlArray)
2018-01-15 20:48:32 +00:00
id='KukaCamBulletEnv-v0',
entry_point='pybullet_envs.bullet:KukaCamGymEnv',
timestep_limit=1000,
reward_threshold=5.0,
Remove duplicate assets and use other assets. Rename classes appropriately for robot-scene-env split refactoring.
2017-08-15 13:31:24 +00:00
)
register(
id='InvertedPendulumBulletEnv-v0',
refactor pybullet/gym to allow instantiating environments directly from a pybullet install: work-in-progress (need to add missing data files, fix paths etc) example: pip install pybullet pip install gym python import gym import pybullet import pybullet_envs env = gym.make("HumanoidBulletEnv-v0")
2017-08-22 07:42:02 +00:00
entry_point='pybullet_envs.gym_pendulum_envs:InvertedPendulumBulletEnv',
Remove duplicate assets and use other assets. Rename classes appropriately for robot-scene-env split refactoring.
2017-08-15 13:31:24 +00:00
max_episode_steps=1000,
reward_threshold=950.0,
)
register(
id='InvertedDoublePendulumBulletEnv-v0',
refactor pybullet/gym to allow instantiating environments directly from a pybullet install: work-in-progress (need to add missing data files, fix paths etc) example: pip install pybullet pip install gym python import gym import pybullet import pybullet_envs env = gym.make("HumanoidBulletEnv-v0")
2017-08-22 07:42:02 +00:00
entry_point='pybullet_envs.gym_pendulum_envs:InvertedDoublePendulumBulletEnv',
Remove duplicate assets and use other assets. Rename classes appropriately for robot-scene-env split refactoring.
2017-08-15 13:31:24 +00:00
max_episode_steps=1000,
reward_threshold=9100.0,
)
register(
id='InvertedPendulumSwingupBulletEnv-v0',
refactor pybullet/gym to allow instantiating environments directly from a pybullet install: work-in-progress (need to add missing data files, fix paths etc) example: pip install pybullet pip install gym python import gym import pybullet import pybullet_envs env = gym.make("HumanoidBulletEnv-v0")
2017-08-22 07:42:02 +00:00
entry_point='pybullet_envs.gym_pendulum_envs:InvertedPendulumSwingupBulletEnv',
Remove duplicate assets and use other assets. Rename classes appropriately for robot-scene-env split refactoring.
2017-08-15 13:31:24 +00:00
max_episode_steps=1000,
reward_threshold=800.0,
)
Improve PyBullet ports of Roboschool envs: fix reset (it kept adding stadium objects, causing slowdown), now reset uses saveState/restoreState and reset becomes a few orders of magnitude faster. Use python -m pybullet_envs.examples.testEnv --env AntBulletEnv-v0 --render=1 --steps 1000 --resetbenchmark=1 Added environments: HumanoidFlagrunBulletEnv-v0, HumanoidFlagrunHarderBulletEnv-v0, StrikerBulletEnv-v0, ThrowerBulletEnv-v0, PusherBulletEnv-v0, ReacherBulletEnv-v0, CartPoleBulletEnv-v0 and register them to OpenAI Gym. Allow numpy/humanoid_running.py to use abtch or non-batch update (setJointMotorControl2/setJointMotorControlArray)
2018-01-15 20:48:32 +00:00
register(
id='ReacherBulletEnv-v0',
entry_point='pybullet_envs.gym_manipulator_envs:ReacherBulletEnv',
max_episode_steps=150,
reward_threshold=18.0,
)
register(
id='PusherBulletEnv-v0',
entry_point='pybullet_envs.gym_manipulator_envs:PusherBulletEnv',
max_episode_steps=150,
reward_threshold=18.0,
)
register(
id='ThrowerBulletEnv-v0',
entry_point='pybullet_envs.gym_manipulator_envs:ThrowerBulletEnv',
max_episode_steps=100,
reward_threshold=18.0,
)
register(
id='StrikerBulletEnv-v0',
entry_point='pybullet_envs.gym_manipulator_envs:StrikerBulletEnv',
max_episode_steps=100,
reward_threshold=18.0,
)
Remove duplicate assets and use other assets. Rename classes appropriately for robot-scene-env split refactoring.
2017-08-15 13:31:24 +00:00
register(
id='Walker2DBulletEnv-v0',
refactor pybullet/gym to allow instantiating environments directly from a pybullet install: work-in-progress (need to add missing data files, fix paths etc) example: pip install pybullet pip install gym python import gym import pybullet import pybullet_envs env = gym.make("HumanoidBulletEnv-v0")
2017-08-22 07:42:02 +00:00
entry_point='pybullet_envs.gym_locomotion_envs:Walker2DBulletEnv',
Remove duplicate assets and use other assets. Rename classes appropriately for robot-scene-env split refactoring.
2017-08-15 13:31:24 +00:00
max_episode_steps=1000,
reward_threshold=2500.0
)
register(
id='HalfCheetahBulletEnv-v0',
refactor pybullet/gym to allow instantiating environments directly from a pybullet install: work-in-progress (need to add missing data files, fix paths etc) example: pip install pybullet pip install gym python import gym import pybullet import pybullet_envs env = gym.make("HumanoidBulletEnv-v0")
2017-08-22 07:42:02 +00:00
entry_point='pybullet_envs.gym_locomotion_envs:HalfCheetahBulletEnv',
Remove duplicate assets and use other assets. Rename classes appropriately for robot-scene-env split refactoring.
2017-08-15 13:31:24 +00:00
max_episode_steps=1000,
reward_threshold=3000.0
)
register(
id='AntBulletEnv-v0',
refactor pybullet/gym to allow instantiating environments directly from a pybullet install: work-in-progress (need to add missing data files, fix paths etc) example: pip install pybullet pip install gym python import gym import pybullet import pybullet_envs env = gym.make("HumanoidBulletEnv-v0")
2017-08-22 07:42:02 +00:00
entry_point='pybullet_envs.gym_locomotion_envs:AntBulletEnv',
Remove duplicate assets and use other assets. Rename classes appropriately for robot-scene-env split refactoring.
2017-08-15 13:31:24 +00:00
max_episode_steps=1000,
reward_threshold=2500.0
)
Improve PyBullet ports of Roboschool envs: fix reset (it kept adding stadium objects, causing slowdown), now reset uses saveState/restoreState and reset becomes a few orders of magnitude faster. Use python -m pybullet_envs.examples.testEnv --env AntBulletEnv-v0 --render=1 --steps 1000 --resetbenchmark=1 Added environments: HumanoidFlagrunBulletEnv-v0, HumanoidFlagrunHarderBulletEnv-v0, StrikerBulletEnv-v0, ThrowerBulletEnv-v0, PusherBulletEnv-v0, ReacherBulletEnv-v0, CartPoleBulletEnv-v0 and register them to OpenAI Gym. Allow numpy/humanoid_running.py to use abtch or non-batch update (setJointMotorControl2/setJointMotorControlArray)
2018-01-15 20:48:32 +00:00
register(
id='HopperBulletEnv-v0',
entry_point='pybullet_envs.gym_locomotion_envs:HopperBulletEnv',
max_episode_steps=1000,
reward_threshold=2500.0
)
Remove duplicate assets and use other assets. Rename classes appropriately for robot-scene-env split refactoring.
2017-08-15 13:31:24 +00:00
register(
id='HumanoidBulletEnv-v0',
refactor pybullet/gym to allow instantiating environments directly from a pybullet install: work-in-progress (need to add missing data files, fix paths etc) example: pip install pybullet pip install gym python import gym import pybullet import pybullet_envs env = gym.make("HumanoidBulletEnv-v0")
2017-08-22 07:42:02 +00:00
entry_point='pybullet_envs.gym_locomotion_envs:HumanoidBulletEnv',
Remove duplicate assets and use other assets. Rename classes appropriately for robot-scene-env split refactoring.
2017-08-15 13:31:24 +00:00
max_episode_steps=1000
)
Split locomotion environments.
2017-08-15 19:33:55 +00:00
register(
Improve PyBullet ports of Roboschool envs: fix reset (it kept adding stadium objects, causing slowdown), now reset uses saveState/restoreState and reset becomes a few orders of magnitude faster. Use python -m pybullet_envs.examples.testEnv --env AntBulletEnv-v0 --render=1 --steps 1000 --resetbenchmark=1 Added environments: HumanoidFlagrunBulletEnv-v0, HumanoidFlagrunHarderBulletEnv-v0, StrikerBulletEnv-v0, ThrowerBulletEnv-v0, PusherBulletEnv-v0, ReacherBulletEnv-v0, CartPoleBulletEnv-v0 and register them to OpenAI Gym. Allow numpy/humanoid_running.py to use abtch or non-batch update (setJointMotorControl2/setJointMotorControlArray)
2018-01-15 20:48:32 +00:00
id='HumanoidFlagrunBulletEnv-v0',
entry_point='pybullet_envs.gym_locomotion_envs:HumanoidFlagrunBulletEnv',
Split locomotion environments.
2017-08-15 19:33:55 +00:00
max_episode_steps=1000,
Improve PyBullet ports of Roboschool envs: fix reset (it kept adding stadium objects, causing slowdown), now reset uses saveState/restoreState and reset becomes a few orders of magnitude faster. Use python -m pybullet_envs.examples.testEnv --env AntBulletEnv-v0 --render=1 --steps 1000 --resetbenchmark=1 Added environments: HumanoidFlagrunBulletEnv-v0, HumanoidFlagrunHarderBulletEnv-v0, StrikerBulletEnv-v0, ThrowerBulletEnv-v0, PusherBulletEnv-v0, ReacherBulletEnv-v0, CartPoleBulletEnv-v0 and register them to OpenAI Gym. Allow numpy/humanoid_running.py to use abtch or non-batch update (setJointMotorControl2/setJointMotorControlArray)
2018-01-15 20:48:32 +00:00
reward_threshold=2000.0
refactor pybullet/gym to allow instantiating environments directly from a pybullet install: work-in-progress (need to add missing data files, fix paths etc) example: pip install pybullet pip install gym python import gym import pybullet import pybullet_envs env = gym.make("HumanoidBulletEnv-v0")
2017-08-22 07:42:02 +00:00
)
update pybullet docs, add getList to pybullet_envs
2017-08-31 16:59:40 +00:00
Improve PyBullet ports of Roboschool envs: fix reset (it kept adding stadium objects, causing slowdown), now reset uses saveState/restoreState and reset becomes a few orders of magnitude faster. Use python -m pybullet_envs.examples.testEnv --env AntBulletEnv-v0 --render=1 --steps 1000 --resetbenchmark=1 Added environments: HumanoidFlagrunBulletEnv-v0, HumanoidFlagrunHarderBulletEnv-v0, StrikerBulletEnv-v0, ThrowerBulletEnv-v0, PusherBulletEnv-v0, ReacherBulletEnv-v0, CartPoleBulletEnv-v0 and register them to OpenAI Gym. Allow numpy/humanoid_running.py to use abtch or non-batch update (setJointMotorControl2/setJointMotorControlArray)
2018-01-15 20:48:32 +00:00
register(
id='HumanoidFlagrunHarderBulletEnv-v0',
entry_point='pybullet_envs.gym_locomotion_envs:HumanoidFlagrunHarderBulletEnv',
max_episode_steps=1000
)
#register(
# id='AtlasBulletEnv-v0',
# entry_point='pybullet_envs.gym_locomotion_envs:AtlasBulletEnv',
# max_episode_steps=1000
# )
def getList():
btenvs = ['- ' + spec.id for spec in gym.envs.registry.all() if spec.id.find('Bullet')>=0]
return btenvs
Reference in New Issue Copy Permalink