AuroraMiddleware/bullet3
1
0
Fork 0
mirror of https://github.com/bulletphysics/bullet3 synced 2024-12-14 05:40:05 +00:00
Code Issues Releases Wiki Activity
8,953 Commits 19 Branches 27 Tags 326 MiB
513e3a7db4
Commit Graph

280 Commits

Author SHA1 Message Date
erwincoumans
024af08320 Added laikago mocap data for a DeepMimic compatible walk cycle 
Added testLaikago.py script to test this mocap data.
 2019-04-04 19:40:21 -07:00
Erwin Coumans
4898887265 temporary disable experimental BulletRobotics examples 2019-03-26 15:07:55 -07:00
Erwin Coumans
2ba8c22397 add actuatornet code 2019-03-26 10:05:45 -07:00
iory
a8146024fd [pybullet/gym/pybullet_envs] Modified register's timestep_limit to max_episode_steps for gym 2019-03-24 03:32:20 +09:00
erwincoumans
9662ae4ba0
Update kuka_diverse_object_gym_env.py 
Fix issue in glob to separate training/test data pattern. See https://github.com/bulletphysics/bullet3/issues/2158
Thank to @TomoyaFukui
 2019-03-14 09:33:48 -07:00
Erwin Coumans
03e897d1f5 replace deep_mimic policies by versions trained using PyBullet, backflip works better now. 2019-02-24 20:12:18 -08:00
Erwin Coumans
6d2ed912f5 Merge remote-tracking branch 'bp/master' 2019-02-18 19:11:03 -08:00
erwincoumans
7e5ac33cf1
Merge pull request #2105 from Somedaywilldo/master 
Add Human3.6M Support for deep_mimic Environment
 2019-02-18 18:07:47 -08:00
erwincoumans
bdf9b10246 more work on pybullet_envs.deep_mimic. 
allow btMultiBody to not wakeup (for some RL experiments)
move deep_mimic motion files to data/motions folder, so we can use the args files unmodified.
 2019-02-18 17:57:02 -08:00
erwincoumans
79a273f644 deep_mimic: add option for spherical joint drive motor, next to stable PD control 
(existing policies won't work with those motors, needs tuning and re-training)
 2019-02-12 20:42:05 -08:00
Somedaywilldo
59d58ce430 add Human3.6M support for deep_mimic env 2019-02-13 00:13:56 +08:00
Erwin Coumans
aaae366a21 towards python 2.x support for deep_mimic (still issue with Logger.print2) 2019-02-11 10:03:26 -08:00
Erwin Coumans
786ddec076 disable GUI in deep_mimic, it slows down too much on Mac OSX 2019-02-11 09:44:59 -08:00
Erwin Coumans
30290dfba1 add missing __init__.py for pip 2019-02-11 09:20:44 -08:00
Erwin Coumans
2728294c53 more fixes in deep_mimic env 2019-02-11 09:15:42 -08:00
Erwin Coumans
1bd201eb43 fixes in PyBullet deep_mimic to allow running in pip version 2019-02-11 08:51:07 -08:00
erwincoumans
d4292fdac3 PyBullet deep_mimic fix computeAngVelRel 2019-02-10 21:37:34 -08:00
erwincoumans
be20ed6aa2 remove obsolete deep_mimic files (use the one in deep_mimic/env) 
add plane_implicit.urdf
 2019-02-10 21:04:28 -08:00
erwincoumans
9bddca873c allow pybullet_envs.deep_mimic.testrl --arg_file run_humanoid3d_backflip_args.txt to perform a backflip. Can only backflip twice, then drops on ground. 
this deepmimic is still very slow, due to slow mass matrix/inverse dynamics computation. once spherical motor drive is enabled, it should be fast(er)
move pd_controller_stable to pybullet_utils for easier re-use
add plane_transparent.urdf to pybullet_data
allow spacebar in keyboardEvents (Windows for now)
 2019-02-10 20:56:31 -08:00
erwincoumans
28c9ea3aad PyBullet: add domino asset and example 2019-02-10 19:44:39 -08:00
Erwin Coumans
2eace2f715 update mpi usage 2019-02-01 19:20:08 -08:00
Erwin Coumans
176edeff1e add mpi_run version (not working yet) 2019-02-01 17:57:31 -08:00
Erwin Coumans
1cef924973 deepmimic fix: reference initialization wasn't spread over all frames 2019-02-01 17:18:13 -08:00
Erwin Coumans
7cc2685faf enable deepmimic training on mac 2019-02-01 16:24:48 -08:00
erwincoumans
63e781b88b PyBullet env for DeepMimic (preliminary work-in-progress) 2019-01-31 21:32:51 -08:00
erwincoumans
5d871806eb enable stable PD plugin in premake4.lua PyBullet 
add a normalize in the setRotation axis/angle to be sure.
Add more code from the DeepMimic project, training doesn't work yet.
 2019-01-31 21:31:26 -08:00
Erwin Coumans
63683e8f02 fix Issue 2039 2019-01-28 16:24:44 -08:00
Erwin Coumans
0818112ede fix some gym envs 2019-01-28 16:21:52 -08:00
erwincoumans
c97d1041ed
Merge pull request #2037 from araffin/update-gym 
Update gym envs
 2019-01-28 10:14:09 -08:00
erwincoumans
505cb15600 proto2csv 2019-01-27 12:11:40 -08:00
erwincoumans
6c2568bf5e PyBullet Minitaur: Fix Minitaur logging (binary file) 
PyBullet Minitaur: Add minitaur_raibert_controller_example.py
 2019-01-27 11:37:35 -08:00
erwincoumans
03549ca7c6 pass force as array instead of value 2019-01-23 17:24:10 -08:00
gaurav pathak
7f73081a89
Added code for issue #2036 
Made necessary changes to ```humanoid_deep_mimic_env.py``` file, setting ```initial_frame``` to new position after simulation steps and setting ```SimTime``` using this new initial frame time
 2019-01-03 18:17:08 -05:00
Antonin RAFFIN
f9fdf89c06 Fix indentation 2018-12-31 12:52:43 +01:00
Antonin RAFFIN
1c61b629e3 Fix formating issues 2018-12-31 12:48:42 +01:00
Antonin RAFFIN
0df3527884 Fix gym deprecation warnings 2018-12-28 14:30:05 +01:00
Erwin Coumans
2e30a9565b add humanoid_deepmimic_gym_env.py for HumanoidDeepMimicBulletEnv-v1 (still untested) 2018-11-25 15:33:28 -08:00
erwincoumans
bb6f4fb17c improve the ARS implementation: add multiprocessing Gym environment stepping, add command-line parameters to resume a policy, --render, --movie, --steps, --env 2018-11-24 11:54:36 -08:00
erwincoumans
ca36a82c62 add deepmimic args file (for testing) 
implement deepmimic reward function (only joint angles/velocities for now)
fix an out-of-date comment, related to contactPoint.m_lateralFrictionInitialized in btSequentialImpulseConstraintSolver.cpp
 2018-11-23 18:32:23 -08:00
erwincoumans
121cdc91b0 more work on PyBullet implementation of DeepMimic humanoid mimic of motion capture. 
b3Quaternion, deal with zero-length axis (in axis,angle constructor)
 2018-11-21 11:09:10 -08:00
erwincoumans
9e99f5cdbc Expose PyBullet.calculateVelocityQuaternion, getAxisAngleFromQuaternion, getQuaternionFromAxisAngle, getDifferenceQuaternion 
Add preparation for DeepMimic humanoid environment, replicating parts of https://github.com/xbpeng/DeepMimic
Loading humanoid.urdf and applying motion action: examples/pybullet/gym/pybullet_envs/mimic/humanoid.py
Loading MotionCapture data: examples/pybullet/gym/pybullet_envs/mimic/motion_capture_data.py
Little test: examples/pybullet/gym/pybullet_envs/mimic/humanoid_test.py
 2018-11-16 17:29:03 -08:00
Erwin Coumans
8b53e47fe8 add simpler ARS implementation, thanks to Alexis Jacq and Hadelin de Ponteves 
(will add save/restore of policy and rendering movies through command-line arguments soon)
 2018-11-02 11:19:46 -07:00
Erwin Coumans
44976780fa remove ARS files 2018-10-29 19:23:54 -07:00
Erwin Coumans
c2b9dc9361 update to CartPoleBulletEnv-v1 and check with latest baselines v0.1.5, works fine. 
(make it more similar to classical control cartpole)
 2018-10-29 16:44:12 -07:00
Erwin Coumans
0511a36a67 add Laikago quadruped model, thanks to Unitree for the cad meshes. 
If you use the Laikago URDF, please add a citation to PyBullet:

@MISC{coumans2018,
author =   {Erwin Coumans and Yunfei Bai},
title =    {PyBullet, a Python module for physics simulation for games, robotics and machine learning},
howpublished = {\url{http://pybullet.org}},
year = {2016--2018}
}
 2018-10-26 18:49:58 -07:00
erwincoumans
b5e475aec3 enable planar reflection in MinitaurGymEnv 
enable follow cam in other Gym locomotion environments
add testing assets for multi-material obj files -> sdf conversion.
Also use ER_NO_SEGMENTATION_MASK flag for TinyRenderer/EGL plugin renderer
 2018-10-14 15:10:19 -07:00
Joshua Aduol
cc4c02edc7 Fixed bugs mistaking forearm_roll_joint for upper_arm_roll_joint 2018-09-09 00:17:06 +01:00
Alexis David Jacq
6adea4964f
forgot one "isAlive" to change 2018-06-16 14:00:05 +02:00
Alexis David Jacq
50b5edd7b5
isDone instead of isAlive 2018-06-16 13:53:05 +02:00
Alexis David Jacq
292a3f1cf6
methode isAlive 
Method def isAlive(self), which defaults to return self._alive < 0, and each environment can override this method (Half Cheetah would implement return False)

(In response to bea468fb93)
 2018-06-15 17:55:20 +02:00
Alexis David Jacq
a57c480f28
Update gym_locomotion_envs.py 
As suggested in https://github.com/bulletphysics/bullet3/pull/1759. The default isDone lets done = alive<0, and a special case is made for halfcheetah, forcing done=False. 
I had to pass the 'alive' condition as an additive parameter of WalkerBaseBulletEnv.
 2018-06-15 17:38:35 +02:00
Erwin Coumans
e24da97140 revert GJK change yet again, some degenerate case / regression 2018-06-12 19:52:38 -07:00
Erwin Coumans
e74f86d148 remove ANYmal meshes 2018-06-11 07:44:05 -07:00
erwincoumans
5ede227e94
Update robot_bases.py 
remove print
 2018-06-05 19:33:28 -07:00
Erwin Coumans
e74cd05e37 add space 2018-05-31 21:21:50 -07:00
Erwin Coumans
76e88dddc5 add space to make conversion work 2018-05-31 18:16:40 -07:00
Erwin Coumans
39808ddfdf allow to run minitaur environment examples from pip version, like this: 
python -m pybullet_envs.minitaur.envs.minitaur_trotting_env_example
python -m pybullet_envs.minitaur.envs.minitaur_reactive_env_example
 2018-05-29 09:28:31 -07:00
Erwin Coumans
efe9161670 fix cmake build 2018-05-26 08:46:12 +10:00
Erwin Coumans
5de53d7355 PyBullet: add ANYmal.py quadruped example, usage: 
pip install pybullet
python -m pybullet_envs.examples.ANYmal
 2018-05-25 07:54:38 +10:00
erwincoumans
f1b4ffaf98
Merge pull request #1699 from benelot/fix-pendula-with-pbclient 
Fix pendulum & manipulator envs with pybullet client
 2018-05-23 08:20:40 +10:00
Benelot
4a16032820 Fix manipulators too. 2018-05-22 18:12:45 +02:00
Benelot
7b219e0ea6 Fix pendula to use the bullet-client. 2018-05-22 16:40:10 +02:00
Erwin Coumans
f5952a73e7 UrdfImporter: Collada DAE, rudimentary support for polylist (for ANYmal), assumes all polygons have 3 vertices 
remove obsolete env from __init__.py
 2018-05-23 00:28:49 +10:00
Erwin Coumans
9a706d9cfd PyBullet: fix Gym envs 2018-05-23 00:24:03 +10:00
Erwin Coumans
a0a1c9347c PyBullet: fix in Gym env 2018-05-23 00:18:17 +10:00
Erwin Coumans
a7a6e20d9c Merge branch 'master' of https://github.com/erwincoumans/bullet3 2018-05-20 14:19:49 +10:00
Erwin Coumans
90939279f9 PyBullet: cmake build, enable DBUILD_PYBULLET_NUMPY=ON 
fix issue with HalfCheetahBulletEnv-v0 in previous commit
 2018-05-20 14:18:53 +10:00
Erwin Coumans
964f52f62d remove unsupported simpleHumanoid.py, use env = gym.make("HumanoidBulletEnv-v0") instead. 2018-05-20 12:35:33 +10:00
Erwin Coumans
0abe4151e5 Fix for 1643, allow to instantiate multiple PyBullet Gym environments (Ant, Humanoid, Hopper, Pendula etc) in the same process (same or other thread). It uses the pybullet_utils.bullet_client to achieve this. 2018-05-18 16:23:54 -07:00
Erwin Coumans
3a4d1799d8 remove "DO NOT SUBMIT" comments, it blocks commits in google3 2018-05-07 17:35:55 -07:00
Jie Tan
20dd3d265b update README.md to include paper link in Arxiv. 2018-04-30 11:09:59 -07:00
Jie Tan
26cebc9aa5 rewrite the README file. 2018-04-26 21:26:08 -07:00
Jie
a375a349ce make sure that the pre-trained galloping and trotting policies work for the minitaur_reactive_env and minitaur_trotting_env environments. 2018-04-24 21:48:27 -07:00
Erwin Coumans
bd0aed3e36 add ARS to train/eval Minitaur 2018-04-11 18:14:56 -07:00
Erwin Coumans
40b285ebfd fixes in minitaur leg mapping 2018-04-11 18:13:46 -07:00
erwincoumans
2bc246b0b1 fixes in relative imports 2018-04-11 18:00:43 -07:00
Erwin Coumans
ad75ea35b3 fix new minitaur envs for newer Gym (_step -> step etc), 
add extra sleep to sync Minitaur walk with RTC clock
 2018-04-11 13:22:25 -07:00
erwincoumans
ac945df07a remote 'render' arg 2018-04-11 10:14:24 -07:00
erwincoumans
698b20938f fix a number of issues in a series of new Minitaur environments 
add them to pybullet_envs through __init__.py
    id='MinitaurReactiveEnv-v0',
    id='MinitaurTrottingEnv-v0',
    id='MinitaurBallGymEnv-v0',
    id='MinitaurStandGymEnv-v0',
    id='MinitaurAlternatingLegsEnv-v0',
    id='MinitaurFourLegStandEnv-v0',

disable reflection of minitaur_four_leg_stand_env, since the floor changes orientation (reflection is a fixed plane with [0,0,1] normal)

from pybullet_envs.minitaur.envs.minitaur_alternating_legs_env import MinitaurAlternatingLegsEnv
from pybullet_envs.minitaur.envs.minitaur_ball_gym_env import MinitaurBallGymEnv
from pybullet_envs.minitaur.envs.minitaur_randomize_terrain_gym_env import MinitaurRandomizeTerrainGymEnv
from pybullet_envs.minitaur.envs.minitaur_reactive_env import MinitaurReactiveEnv
from pybullet_envs.minitaur.envs.minitaur_stand_gym_env import MinitaurStandGymEnv
from pybullet_envs.minitaur.envs.minitaur_trotting_env import MinitaurTrottingEnv
from pybullet_envs.minitaur.envs.minitaur_four_leg_stand_env import MinitaurFourLegStandEnv
 2018-04-11 10:09:03 -07:00
erwincoumans
c2869e0a3c enable planar reflection in minitaur_gym_env.py and locomotion scenes 2018-04-11 08:50:29 -07:00
Erwin Coumans
7fa5e91cbd replace green stadium with blue checkered Bullet plane 
add missing files to pybullet setup.pu
 2018-04-10 19:55:26 -07:00
erwincoumans
5ff2a1d6c8
Update env_bases.py 2018-04-02 22:31:22 -07:00
erwincoumans
14d37ecb43 first pass of updated minitaur quadruped environment 2018-03-31 21:15:27 -07:00
erwincoumans
d69fefd7e9
Merge pull request #1559 from shubhtuls/patch-1 
minor bugfix in image reshaping in kukaGymEnv.py
 2018-03-08 15:16:30 -08:00
Erwin Coumans
8ba9644e78 fix __init__.py 2018-02-21 10:47:24 -08:00
Erwin Coumans
67318f3323 fix __init__.py 2018-02-21 10:45:06 -08:00
Erwin Coumans
a3c2ce1cef add __init__.py for pybullet_envs.prediction 2018-02-21 10:06:46 -08:00
erwincoumans
d3fe21ef17 Add option GUI_MAIN_THREAD for running OpenGL in the main thread (Python thread). This already happens on Mac OSX by default. 
Pass all argc/argc from options to InProcessPhysicsClientSharedMemoryMainThread
 2018-02-14 16:49:44 -08:00
erwincoumans
5e08d6d334 small test to create / simulate random box stacks 2018-02-14 15:37:50 -08:00
Shubham Tulsiani
5fdd7ed3b4
minor bugfix in image reshaping in kukaGymEnv 
Reshaped image array should be of size (RENDER_HEIGHT, RENDER_WIDTH, 4) instead of (RENDER_WIDTH, RENDER_HEIGHT, 4).
 2018-02-11 21:47:57 -08:00
erwincoumans
19231434c2 gym env backward compatibility due to gym 0.9.6 breaking changes 2018-02-06 10:20:41 -08:00
Sam Wenke
ad3c236bfd pybullet support for gym.Env, including v0.9.x 2018-02-04 11:35:47 -05:00
erwincoumans
13d67e8fc3 add bicycle resources and testBike.py script (use python -m pybullet_envs.examples.testBike after pip install pybullet) 
center shadow around camera target
 2018-01-31 22:17:15 -08:00
Erwin Coumans
cb0e740422 fix error in __init__.py 2018-01-27 10:25:26 -08:00
Erwin Coumans
2b5be5ca8a Merge remote-tracking branch 'bp/master' 2018-01-26 17:36:45 -08:00
erwincoumans
688d750d8a
Merge pull request #1531 from AndreaIncertiDelmonte/issue-1525 
Override of the reset() public funciton in child classes
 2018-01-26 13:10:47 -08:00
cedspam
3011cf1130
avoid errors in some ide docs system 
there are some error linked to repeated import of this module in spyder, an env with the same id cannot be registered again
 2018-01-26 09:29:39 +00:00
AndreaIncertiDelmonte
b908b147a6 Overwrite of the reset() public funciton in child classes 2018-01-26 10:09:14 +01:00
erwincoumans
0ce40b70c0
Merge pull request #1523 from kimhc6028/master 
[fixed] gym_manipulator_envs typo
 2018-01-25 13:36:42 -08:00
Erwin Coumans
c6fefb1018 remove useless gym env 2018-01-25 12:11:22 -08:00
Erwin Coumans
c4a046f16c remove baselines dependency (not needed, since it is a random policy) 2018-01-25 07:34:00 -08:00
erwincoumans
5d45fd5ca5 PyBullet: register KukaDiverseObjectGrasping-v0 to Gym. 2018-01-24 18:27:03 -08:00
erwincoumans
4ea406c73c PyBullet: improve examples/pybullet/gym/pybullet_envs/bullet/kuka_diverse_object_gym_env.py 
(thanks to Deirdre Quillen for the environment)
Extend repeat, so gripper reaches the tray bottom.
Fix near plane so Z-Buffer is visible. Add sleep in return motion, in gui mode.
 2018-01-24 18:21:59 -08:00
erwincoumans
72ec3effe6 Merge remote-tracking branch 'bp/master' 2018-01-24 18:11:33 -08:00
erwincoumans
583b23ea40 PyBullet Humanoid Flagrun Harder: disable collision on target red sphere 2018-01-24 18:02:24 -08:00
erwincoumans
aba4671358 Humanoid Flagrun Harder, PyBullet version (request from Danijar) 
Fix duplicate ground in all Roboschool converted scenes.
Fix bug in all PyBullet-Roboschool converted Humanoid environments (action needs to be clipped to [-1,1] range)
 2018-01-24 17:53:40 -08:00
Erwin Coumans
b8362fff94 PyBullet pybullet_envs: fix issue with Humanoid environments (excessive forces, due to lack of action clamping) 
PyBullet pybullet_envs: use 1./60 sleep, add enjoy_TF_HumanoidFlagrunHarderBulletEnv_v1_2017jul.py
 2018-01-24 16:02:19 -08:00
kimhc6028
0d07ae03b7 [fixed] gym_manipulator_envs typo 2018-01-23 18:22:53 +09:00
Erwin Coumans
69007dcc61 fix MJCF loading (use self-collision) 2018-01-20 12:07:35 -08:00
deq2
e30f2624f9 Added kuka environment with multiple diverse objects. (#1508) 
* Added new diverse object environment
 2018-01-18 08:49:00 -08:00
Erwin Coumans
c6a350b807 BulletInverseDynaimcs: re-applied fix 
PyBullet: increase humanoid training from 30 to 300M (need to figure out the right hyperparameters...)
 2018-01-16 21:54:46 -08:00
erwincoumans
055930817e apply a maximum velocity for the KUKA arm, otherwise motion can become very unrealistic 2018-01-16 10:17:57 -08:00
erwincoumans
851ca5bfb3 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 12:48:32 -08:00
erwincoumans
22b4809891 leave m_useGlobalVelocities to false, until enabled, for backward compatibility 
use URDF_GLOBAL_VELOCITIES_MB flag in PyBullet loadURDF.
fix robot_bases.py due to new fields in getJointInfo.
backward compabitibility: BulletMJCFImporter, keep creating btMultiSphereShape for MJCF capsules with fromto, instead of shifted btCapsuleShapeZ, unless if CUF_USE_IMPLICIT_CYLINDER is used.
 2018-01-10 11:16:50 -08:00
Erwin Coumans
c88132b80f fix issue in pybullet_envs.agents 
bump up to pybullet 1.6.9
 2017-11-18 17:07:27 -08:00
Erwin Coumans
f483ccd8ef remote org configs 2017-11-16 16:49:25 +00:00
Erwin Coumans
7b030426c1 add a temp copy of TF agents (until the API stops changing or configs.py are included) 2017-11-16 16:47:14 +00:00
erwincoumans
6be7e34dd6 pybullet vr_kuka_setup.py: add a gear joint, to keep the gripper centered, 
vr_kuka_control.py: control all joints, use analogue button to close gripper
remove some debug warnings/prints
pybullet, avoid crash in changeUserConstraint if not passing a [list]
allow some gym environments (pybullet_pendulum and locomotors) to re-use an existing physics client connection.
 2017-11-12 10:36:30 -08:00
Erwin Coumans
4798d66f19 fixes in kuka/racecar environment. kuka still doesn't train well, work-in-progress 2017-11-01 18:06:47 -07:00
Erwin Coumans
e4f58ddc33 fix typo, add train_ppo kuka task 2017-10-31 16:33:13 -07:00
Erwin Coumans
3709e68645 fix discrete versions of kuka grasping environment 
should fix issue 1386
 2017-10-31 16:14:20 -07:00
Erwin Coumans
33d32d7ab4 add continuous version of kuka cam gym env test 2017-10-31 16:10:02 -07:00
Erwin Coumans
4e2dfbe58e 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 16:05:52 -07:00
Erwin Coumans
55f5e52ecd add continuous versions of kukaGymEnv, kukaCamGymEnv, racecarZEDGymEnv etc. 
should be trainable with PPO or evolution strategies (ES) now
 2017-10-31 15:50:34 -07:00
Erwin Coumans
9385f36505 add cartpole_bullet, should fix issue 1403 2017-10-25 08:37:01 -07:00
Erwin Coumans
d18bfec1a9 API changes in TensorFlow Agents 
See 164f620326 (diff-82bda908f176861c20aadc1d017c2527)
 2017-10-20 20:32:49 -07:00
Erwin Coumans
91d164d886 remove obsolete CartPoleBulletEnv 2017-10-09 09:44:54 -07:00
Erwin Coumans
0948ce5984 import missing functools 2017-09-28 10:42:02 -07:00
Erwin Coumans
345836d670 revert to original agents train/visualize scripts, but using pybullet envs in configs.py (agents lacks a convenient way to extend environments) 2017-09-28 10:19:41 -07:00
yunfeibai
85be3b43dd Merge remote-tracking branch 'upstream/master' 2017-09-27 15:20:11 -07:00
yunfeibai
113e103bc2 Modify the joint damping input check, and output error message when it doesn't pass the size check. Modify the kuka grasping example accordingly. Setting joint damping in this example was a no-op before. 2017-09-27 15:18:08 -07:00
Erwin Coumans
4dca5b3ef5 fix typos in agents config 2017-09-27 12:17:59 -07:00
Erwin Coumans
a716752915 add __init__.py file in agents folder 2017-09-27 10:30:05 -07:00
Erwin Coumans
5082d6af15 bump up pybullet version, instruction to visualize agents ppo: 
python -m pybullet_envs.agents.visualize_ppo -config=pybullet_pendulum --logdir=pendulum/20170927T101514-pybullet_pendulum/ --outdir=vid

tensorboard --logdir=20170927T101514-pybullet_pendulum --port=2222
 2017-09-27 10:26:45 -07:00
Erwin Coumans
e4a3b3fe38 add TensorFlow Agents PPO training script for various pybullet environments: 
example:

python -m pybullet_envs.agents.train_ppo --config=pybullet_pendulum --logdir=pendulum
 2017-09-27 10:20:38 -07:00
Erwin Coumans
c37919604f add minitaur environment randomizer, enable it by default 
this improves PPO training ( more stable)
 2017-09-27 09:07:21 -07:00
Erwin Coumans
f4b28dddce fix rendering ("rgb_array") for pendulums/pendula 2017-09-15 16:33:22 -07:00
Erwin Coumans
81bf709ad7 Merge branch 'master' of https://github.com/erwincoumans/bullet3 2017-09-15 13:12:06 -07:00
Erwin Coumans
3d7ee317db fix pybullet gym environments: InvertedDoublePendulumBulletEnv-v0, 
InvertedPendulumBulletEnv-v0, InvertedPendulumSwingupBulletEnv-v0

motors were not disabled, extra gains applied etc.
 2017-09-15 13:09:05 -07:00
erwincoumans
ffb05d74da Merge pull request #1308 from benelot/fix-gyms 
Fix gyms to make them work just like robotschool
 2017-09-14 07:40:51 -07:00
Benelot
3f57fb655a Enable randomized reset for Humanoid. 2017-09-14 12:53:12 +02:00
Benelot
2e8a86462f Remove tweaks added by me but not part of roboschool. 2017-09-14 12:52:21 +02:00
Erwin Coumans
340a8f4704 add sigaction handler to Example Browser, to always shutdown shared memory 
make GUI_SERVER more reliable
next attempt to connect to SHARED_MEMORY in Gym envs, if available, before DIRECT/GUI
allow software rendering fallback, even if ER_BULLET_HARDWARE_OPENGL is chosen in getCameraImage
 2017-09-13 13:30:16 -07:00
Erwin Coumans
c250a5f0b9 re-enable shared memory connection for pybullet Gym envs (with fallback to GUI or DIRECT) 
suppress shared memory connection warnings
add fallback from ER_BULLET_HARDWARE_OPENGL to TinyRenderer if not available
 2017-09-13 09:56:39 -07:00
Erwin Coumans
3891602784 enable continuous action space for racecarZEDGymEnv 
disable SHARED_MEMORY connection (it was just some debug test)
 2017-09-11 17:34:06 -07:00
Erwin Coumans
105c2c948f add simple 'runServer.py' script, that lets you run a GUI shared memory server to connect to 2017-09-11 09:38:55 -07:00
Erwin Coumans
f38b2cf14d use SHARED_MEMORY as first connection attempt, to make visualization/OpenGL rendering easier 2017-09-11 09:23:14 -07:00
Erwin Coumans
ee082396d1 fixes in libdl/DL cmake 
minor tweaks in pybullet OpenAI gym locomotion envs (connect to shared memory before going GUI/DIRECT)
bump up pybullet version to 1.3.6
 2017-09-10 16:22:45 -07:00
Erwin Coumans
466c853489 add testMJCF.py script, to visualize MJCF file 2017-09-10 10:45:38 -07:00
Erwin Coumans
125380ce15 revert feet related reward in ant 2017-09-09 22:35:48 -07:00