v3.0.1

What's Changed

• [Feature] Gymnasium >=1.1 support by @StoneT2000 in #1094
• [BugFix] Fix shape issue with cached resets when using an existing dataset of states by @StoneT2000 in #1355
• [Bugfix] Correctly add MultiAgent sensor configs to env by @fheller1 in #1357
• [Docs] Update installation docs with uv and latest mac sapien version by @StoneT2000 in #1370
• update task dexterity/rotate_single_object_in_hand by @Jianshu-Hu in #1377
• fixed typo by @jack-sherman01 in #1389
• [Docs] Fix PlugCharger-v1 dense reward label by @Lidang-Jiang in #1402
• [BugFix] Fix AttributeError with gymnasium >= 1.0 in replay_trajectory by @Lidang-Jiang in #1403
• Fix potential XXE and environment signature mismatch by @RinZ27 in #1392
• [Docs] Update replay_for_il_baselines.sh to optionally use MS_ASSET_DIR by @StoneT2000 in #1409
• [Docs] Remove mention of beta, prepare for 3.0.0 release by @StoneT2000 in #1410

New Contributors

• @fheller1 made their first contribution in #1357
• @Jianshu-Hu made their first contribution in #1377
• @jack-sherman01 made their first contribution in #1389
• @Lidang-Jiang made their first contribution in #1402
• @RinZ27 made their first contribution in #1392

Full Changelog: v3.0.0b22...v3.0.1

v3.0.0b22

What's Changed

• Remove redundant num_actors definition by @Gamenot in #1056
• [BugFix] Fix some NaN issues with physx backend for fixed tendon modelled robots. by @StoneT2000 in #1080
• Update doc to use correct vulkan package on ubuntu by @akashsharma02 in #1086
• [BugFix]: Fetch IK bug fix on GPU sim by @arth-shukla in #996
• [BugFix] Register missing asset download IDs for the TurnFaucet task by @StoneT2000 in #1092
• Add panda_wristcam support for Behavioral Cloning RGBD by @chenyenru in #814
• [BugFix] Stop warnings for fetch robot mimic controller by @StoneT2000 in #1103
• [Documentation] Update benchmark docs with new numbers and files by @StoneT2000 in #1104
• [BUG] Removed empty dense reward function from PlugCharger-v1 by @jstmn in #1096
• [Feature] Support the widowxai robot and provide RL baseline for cube picking, and update task docs by @hu-po in #904
• [Feature] Sim2Real Tooling and Tutorials, LeRobot hardware support by @StoneT2000 in #942
• [Feature] Support different robot colors / robot color domain randomization for SO100GraspCube sim2real by @StoneT2000 in #1132
• [Feature] Support CPU simulation only and fallback options if usable device is not found for rendering by @StoneT2000 in #1133
• [BugFix] Add fix for mac devices not being recognized as being capable of rendering by @StoneT2000 in #1137
• Added ActionRepeatWrapper to wrappers imports by @AlexandreBrown in #1141
• Fixed action repeat diagram not showing online by @AlexandreBrown in #1142
• vectorenv: relax env type annotation to gymnasium.Env by @songyuc in #1144
• [BugFix] fix bug with environments inheriting base digital twin env using certain rgb overlay mode not releasing memory by @StoneT2000 in #1158
• Fix partial reset support for digital twin tasks by @zanghz21 in #1155
• [BugFix] Unified get / set controller state function across multi agent and base agent, fix some issues with recording trajectories by @StoneT2000 in #1161
• [Feature] support setting individual joint qpos values. by @StoneT2000 in #1162
• Stackpyramid Task with Motionplanning Support by @chenyenru in #1088
• [Feature] Add StackPyramid-v1 in motionplanning and il benchmark replay script by @chenyenru in #1172
• Add clamping to [-1, 1] for avoid nans in matrix_to_euler conversion by @tylerlum in #1090
• [Feature] Support set drive targets for individual joint by @StoneT2000 in #1180
• [Feature] support more flexible link merging and update docs by @StoneT2000 in #1191
• [BugFix] fix joint set/get of drive position/velocity targets by @StoneT2000 in #1192
• Add mimic dictionary for xarm6_robotiq by @dwaitbhatt in #1199
• [Feature] Cached resets, fix a few minor reset mask bugs, remove some default dict args by @StoneT2000 in #1203
• [Docs] Fix the documentation around APIs by @StoneT2000 in #1204
• [Feature] Sapien 3 new linux pkg by @StoneT2000 in #1194
• [Feature] Better Multi-GPU Support by @StoneT2000 in #1160
• fix: update root link retrieval by @lvjonok in #1212
• CachedResetWrapper Bug Fix by @ywchoi02 in #1215
• [BugFix] fix dof value in str representation of controllers by @StoneT2000 in #1226
• Fix #1145: Remove unused class attribute cube_half_size by @songyuc in #1229
• [Feature] Faster IK based controllers using levenberg-marqurdt and align pd ee pos controllers in gpu/cpu sims by @StoneT2000 in #1213
• [Docs] Fix typo in domain randomization docs by @StoneT2000 in #1235
• [BugFix] Fix SO100 default qpos and basic motion planning script by @StoneT2000 in #1242
• [Bug Fix]Scene_Builder overriding with new registration by @tkj2022 in #1244
• [Bug Fix] Fixed bug that caused drop in TDMPC2 evaluation performance by @t-sekai in #1240
• [BugFix] Fix setters for root linear and angular velocities by @paLeziart in #1248
• [BugFix] Update SAC baseline to have deterministic training runs by @StoneT2000 in #1258
• Fix escape sequence warnings by @reeceomahoney in #1262
• [Feature]: Teleop scripts for xarm6 and SO100 + creation of base_motionplanner to abstract out reptitive MP code by @cmcgartoll in #1245
• [Docs] Licensing by @StoneT2000 in #1269
• [BugFix] Fix joint pose by @StoneT2000 in #1270
• [BugFix] Fix bug with cached reset wrapper with counting the number of cached resets for user defined reset states by @StoneT2000 in #1273
• Add missing device=self.device in actor.py, base.py, link.py --- allows for torch.set_default_device("cuda") by @jstmn in #1278
• Resolve #1283: correct example for vectorized eval on CPU by @sfetzel in #1289
• fixed torch.from_numpy keyword arg in base.py by @eyerisshe in #1291
• [Bugfix] Fix vector env shutdown: implement close_extras in ManiSkillVectorEnv (prevents double-close at teardown) by @skyswordw in #1281
• [Feature] Relax numpy requirement to permit numpy > 2.0 by @StoneT2000 in #1268
• [Feature] Add new dextrous hand tasks, more floating dextrous hand examples, support optionally saving during env resets for record episode wrapper by @liuyulinn in #1297
• [BugFix] fix bug with replay trajectory assuming sim config has some keys by @StoneT2000 in #1306
• [Feature] Fix all old typings that used typings package and replace with built-ins by @StoneT2000 in #1307
• Add missing self.device arguments by @jstmn in #1305
• fix issue #1304 - change default task robot by @liuyulinn in #1310
• [BugFix] Fix InsertFlower-v1 task to be gpu sim compatible by @StoneT2000 in #1311
• Fix typo: 'unecesary' → 'unnecessary' in BaseEnv docstring by @songyuc in #1312
• Fix article: 'a object' → 'an object' in simulation_101.md by @songyuc in #1314
• [Docs] Fix hopper hop success conditions by @Arleee1 in #1322
• Adding XLeRobot as officially supported robot by @Vector-Wangel in #1293
• [BugFix] Replace deprecated pynvml with nvidia-ml-py by @skyswordw in #1316
• fix missing right bracket by @sszxc in #1332
• Add ManiSkill to LeRobot dataset converter by @dancher00 in #1174
• [Feature] Update demo robot script to use tyro by @StoneT2000 in #1339
• fix: correct torch device creation to use specified CUDA device ID by @fffffq00 in #1346

New Contributors

• @Gamenot made their first contribution in #1056
• @akashsharma02 made their first contribution in #1086
• @hu-po made their first contribution in #904
• @zanghz21 made their first contribution in #1155
• @tylerlum made their first contribution in #1090
• @lvjonok made their first contribution in #1212
• @tkj2022 made their first contribution in #1244
• @paLeziart made their first contribution in #1248
• @reeceomahoney made their first contribution in #1262
• @cmcgartoll made their first contribution in #1245
• @sfetzel made their first contribution in https://github.com/haosul...

v3.0.0b21

What's Changed

• [Feature] Initial mac support by @StoneT2000 in #941
• [Feature] nightly builds by @StoneT2000 in #947
• [BugFix] fix profiling code for benchmarking for macos platform and update docs with some limitations of macos sapien by @StoneT2000 in #949
• [BugFix] Fix bug where physx_cpu backend permitted num_envs > 1 by @StoneT2000 in #951
• Update xarm6_robotiq to have a more natural rest position by @dwaitbhatt in #929
• [BugFix] Fix render camera local poses not invalidating cache by @StoneT2000 in #974
• [Docs] Update outdated docs function for state registry by @StoneT2000 in #980
• [Feature] More advanced mimic controller configurations + floating inspire hand implementation by @StoneT2000 in #981
• Adding missing .cpu() calls in stack_cube.py, peg_insertion_side.py by @jstmn in #976
• [Feature] Auto generated robot documentation + initial quality ratings by @StoneT2000 in #984
• [Feature] Pretty printed controller repr by @StoneT2000 in #985
• [Feature] Lerobot hardware support + SO100 and koch modelled arms by @StoneT2000 in #986
• [Docs] fix robot docs links and fix incorrect controller naming of koch arm by @StoneT2000 in #988
• [Feature] SO100 PickCube Sim Test by @StoneT2000 in #995
• [Feature] Better so100 model by @StoneT2000 in #997
• [Feature] Add SO100 convex meshes to support motion planning and example convex mesh generation script by @StoneT2000 in #1008
• Enhance SAC with Mixture-of-Expert and BEE Operator for Improved Stability and Performance by @XuGW-Kevin in #788
• Fixed PushCube policy not penalized for holding cube above target by @AlexandreBrown in #1022
• [BugFix] fix bug for cpu sim replay by @StoneT2000 in #1034
• [Feature] Support motionplanning solutions for xarm6_robotiq by @dwaitbhatt in #936
• [BugFix] Update get collision mesh functions by @StoneT2000 in #1035
• Added Gym Env Action Repeat Wrapper by @AlexandreBrown in #832
• [Docs] update citation with RSS 2025 paper details by @StoneT2000 in #1039
• [BugFix] Links in CONTRIBUTING.md by @suessmann in #1041
• Fixed action repeat image not shown by @AlexandreBrown in #1046
• [Feature] Expose API for incoming link joint forces on CPU and GPU sim by @StoneT2000 in #1061
• [Docs] Add missing imports in setup.md documentation files by @StoneT2000 in #1062
• [BugFix] Zero out articulation qf values upon gpu sim setup by @StoneT2000 in #1063
• [BugFix] Fix bug with multi-agent envs not working by @StoneT2000 in #1064
• Adding Panda motionplanning solution for PlaceSphere-v1 environment by @jstmn in #1068
• Updated Pull Cube Default Camera Config To Show Cube & Target by @AlexandreBrown in #1069
• [BugFix] Update xarm6 asset download link by @StoneT2000 in #1078

New Contributors

• @dwaitbhatt made their first contribution in #929
• @jstmn made their first contribution in #976
• @XuGW-Kevin made their first contribution in #788
• @suessmann made their first contribution in #1041

Full Changelog: v3.0.0b20...v3.0.0b21

v3.0.0b20

What's Changed

• [BugFix] fix framestack wrapper not accepting env_idx=0...num_envs (resetting all envs and not a partial reset). by @StoneT2000 in #888
• Fixed video links for drawing task docs by @arnavg115 in #890
• [BugFix] create numpy array first then torch tensor for some properties by @StoneT2000 in #894
• [Feature] Updated TD-MPC2 Baseline to Support 128x128 RGB with Extra State Data by @t-sekai in #903
• [Feature] update documentation on actor/link domain randomization and make it possible to domain randomize robots/agents by @StoneT2000 in #910
• Fix: viewer + rgbd obs by @arth-shukla in #897
• [BugFix]: ycb checksum by @arth-shukla in #918
• [BugFix] fixed drawing mp solution imports by @arnavg115 in #916
• Update interactive_panda.py by @Ethan-Chen-plus in #917
• fix gpu sim demo, mem type is None by @SajjadPSavoji in #924
• [BugFix] Fix bug where articulations built with initial poses not at 0 would have incorrectly generated collision meshes + deprecate old get articulation meshes function by @StoneT2000 in #927

New Contributors

• @Ethan-Chen-plus made their first contribution in #917
• @SajjadPSavoji made their first contribution in #924

Full Changelog: v3.0.0b19...v3.0.0b20

v3.0.0b19

What's Changed

• [Feature] Upgrade pytorch_kinematics by @StoneT2000 in #854
• Fixed final_info having wrong elapsed_steps value after reset and final_observation being wrong on CUDA by @AlexandreBrown in #828
• Fix agents' qpos type to be floats by @shaido987 in #856
• [Docs] Update docs on IL baselines regarding standard IL datasets and how to replay them, remove old examples.sh files by @StoneT2000 in #861
• YCB, RCAD scene builder updates, demo_manual_control by @arth-shukla in #859
• [BugFix] update the checksum for ycb dataset by @StoneT2000 in #865
• [BugFix] Fix slow CPU sim and fix the CPU sim benchmark code to be faster by @StoneT2000 in #870
• [Feature] Added RGB DP baseline and DrawTriangle/SVG task by @StoneT2000 in #876
• [Feature] Update PPO baselines for pick single YCB and remove wandb-entity CLI arguments by @StoneT2000 in #863
• Fix crash in CPU backend for multi-body Actors by @shaido987 in #880

New Contributors

• @AlexandreBrown made their first contribution in #828
• @shaido987 made their first contribution in #856

Full Changelog: v3.0.0b18...v3.0.0b19

v3.0.0b18

What's Changed

• [BugFix] Fix off by one error in saving GPU replayed trajectories by @StoneT2000 in #842

Full Changelog: v3.0.0b17...v3.0.0b18

v3.0.0b17

What's Changed

• [Feature] Support pd_joint_pos to pd_ee_pose controller action conversion. Fix bug where motion planning demos were using GPU sim backend to replay by default and weren't working. by @StoneT2000 in #834
• [Feature] Handle more flexible observation modes, support users to check if ground truth state data is requested by @StoneT2000 in #835
• [BugFix] Fix initial poses of static objects in AI2THOR scenes to be not zero in CPU and GPU sim. by @StoneT2000 in #836
• [Feature] Remove extra tensor init in PPO code. by @StoneT2000 in #837
• [BugFix] Fix bug where replay trajectory tool would use torch tensor seeds and try to save them in JSON by @StoneT2000 in #840

Full Changelog: v3.0.0b16...v3.0.0b17

v3.0.0b16

What's Changed

• [Docs/BugFix] Fix bug with sac replay buffer incorrectly filling up after max and add new PPO baselines for many tasks by @StoneT2000 in #731
• Fix documentation for torch_deterministic parameter in ppo.py by @songyuc in #745
• [Feature] Support applying external forces on objects by @StoneT2000 in #752
• Adding Action Chunking with Transformers (ACT) to baselines by @ywchoi02 in #640
• [BugFix] action conversion to end-effector target control having shaky behavior by @StoneT2000 in #756
• [BugFix] fix bug with gpu benchmarking script when running maniskill environments due to missing property by @StoneT2000 in #762
• [Feature] Batched and simpler API for setting locked motion axes by @StoneT2000 in #763
• [Feature] support set mass on GPU by @StoneT2000 in #792
• fixing minor typos in docs by @vncntt in #801
• Remove unnecessary conversion of info to NumPy (#791) by @hesic73 in #796
• [BugFix] Fixed dtype error when running TD-MPC2 on anymalc-reach by @t-sekai in #808
• ManiSkill-HAB to docs by @arth-shukla in #759
• [Feature] Cleanup and standardize the imitation learning baselines and datasets by @StoneT2000 in #811
• [Feature] cleaner baseline scripts for ACT baseline and DP by @StoneT2000 in #815
• [BugFix] Improve PPO default parameters and fix bug with lazy init sensors/cameras by @StoneT2000 in #816
• [Docs] update gallery with more videos by @StoneT2000 in #819
• [Feature] Support state+<visual_textures> mode by @StoneT2000 in #821
• Fix typos by @vncntt in #802
• Fix: vector-gymnasium-wrapper -- fixing 'call' method -- getattr(self.env...) … by @imishani in #783
• [Feature] Support XArm6 with robot iq gripper by @StoneT2000 in #820
• [BugFix] Fix usage of the MS_ASSET_DIR variable by @StoneT2000 in #824

New Contributors

• @songyuc made their first contribution in #745
• @ywchoi02 made their first contribution in #640
• @vncntt made their first contribution in #801
• @imishani made their first contribution in #783

Full Changelog: v3.0.0b15...v3.0.0b16

v3.0.0b15

What's Changed

• [BugFix] Fix bug with record episode wrapper for single CPU envs by @StoneT2000 in #723
• [Feature] Substep Rendering by @StoneT2000 in #724

Full Changelog: v3.0.0b14...v3.0.0b15

v3.0.0b14

What's Changed

• [Version] Version Bump to b14 by @StoneT2000 in #721
• [BugFix] Fix bug with missing init py file in panda stick motion planning folder

Full Changelog: v3.0.0b13...v3.0.0b14