Releases · RoboVerseOrg/RoboVerse

We have a major update on metasim, we use new version number 1.0.0 to track. Now all the basic functions are supported and some of useful APIs are still under active development.

The features we support / developing are as follow:

Core feature

✅ Done

#	Feature
1	All get started
2	replay demo / collect demo
3	RL training (All sim?)
4	get traj file / auto download assets
5	motion planning
6	task register
7	gym.make
8	isaacsim
9	domain randomization: infra & test
10	RL infra standardization
11	IL infra standardization
12	query
13	doc (basically)
14	example tasks for RL (Reach, Humanoidbench)
15	example tasks for IL (Maniskill, libero, rl bench)
16	Dex Hand
17	example tasks & assets (only metasim folder required)
18	list-tensor auto convert
19	asset converter
20	camera: multi, mount

❌ Developing / Fixing / Need Testing

#	Feature
1	isaaclab repair
2	VLA infra & algo
3	EE / Torque / PD control support
4	Unit test
5	Robot Config Class
6	state-based replay
7	Render: advance? light? scene? hybrid?
8	multi robots
9	control infra
10	pip install & release
11	More Scripts, Teleoperate?
12	set_dof_target align for mujoco and genesis

What's Changed

fixed pybullet robot always fix_base_link bug by @boshi-an in #18
Dev/robots doc by @boshi-an in #19
added scenes list by @boshi-an in #20
Dev/robots doc by @boshi-an in #21
Dev/robots doc by @boshi-an in #23
added docs by @boshi-an in #24
update asset converter docs by @charlietcheng in #25
fix: Resolve the Isaac Gym error related to loading robots when fix_base_link == False. by @EmptyBlueBox in #27
added the paper link in the GitHub README by @sjtuyinjie in #28
Pull/29/head by @Fisher-Wang in #30
MuJoCo and IsaacGym Infra Update by @EmptyBlueBox in #52
IsaacGym and IsaacLab Infra Change by @EmptyBlueBox in #59
Fix: Resolve the errors of using the evaluation scripts in roboverse_learn by @szang18 in #61
Dev/states as obs by @Fisher-Wang in #65
[docs/fix] Add clone project to installation.rst by @15101051 in #68
Clean up training DP and ACT, with better docs by @dylan-goetting in #74
Dev/hf auto download by @Fisher-Wang in #78
fix joint name extraction error in mujoco.py by @WeisonWEileen in #84
add device property for handlers by @Fisher-Wang in #88
delete get_observation() method from handler by @Fisher-Wang in #89
updated RL infra by @charlietcheng in #86
Tensorize states by @Fisher-Wang in #92
HumanoidBench speed up ~30x by @geng-haoran in #96
Originally PR 107 by @Fisher-Wang in #108
Quickfix/body and device by @Fisher-Wang in #109
fix minimal test script to fit tensorized states by @Fisher-Wang in #111
quickfix: isaacgym get camera images by @Fisher-Wang in #112
Dev/organize get object joint names by @Fisher-Wang in #114
Fix/bug97 2 by @Fisher-Wang in #116
Dev/optimize mujoco states indexing by @Fisher-Wang in #119
fix doc warning by @Fisher-Wang in #120
Update humanoidbench training code by @EmptyBlueBox in #122
Update humanoidbench training code by @EmptyBlueBox in #125
kinova + 2f85 random action success by @Fisher-Wang in #132
Bugfix/isaaclab multienv state by @Fisher-Wang in #134
infra: add PrimitiveFrameCfg; isaaclab: support it by @Fisher-Wang in #135
infra: add contact force sensor cfg; isaaclab: support it by @Fisher-Wang in #137
Dev/clean math util quat inv by @Fisher-Wang in #140
add camera poses to states; isaaclab: support it by @Fisher-Wang in #141
[pre-commit.ci] pre-commit autoupdate by @pre-commit-ci[bot] in #144
Feature/fix102 by @boshi-an in #131
fix evaluation bug of humanoidbench by @WeisonWEileen in #133
infra script: fix teleop_keyboard to use latest tensor states by @Fisher-Wang in #136
Simpler env move near by @boshi-an in #91
Isaaclab ShaderFixer also repair opacity_texture by @Fisher-Wang in #154
Feature/update clean usd script by @Fisher-Wang in #155
run humanoid parrallel training by @WeisonWEileen in #162
Revert "run humanoid parrallel training" by @WeisonWEileen in #163
[pre-commit.ci] pre-commit autoupdate by @pre-commit-ci[bot] in #160
[refactor] auto download by @Fisher-Wang in #165
isaaclab support setting actuator stiffness and damping by @Fisher-Wang in #152
[refactor] remove rootutils by @Fisher-Wang in #166
Doc/checker cfgs by @Fisher-Wang in #167
[refactor] remove mjcf_path from PrimitiveCubeCfg by @Fisher-Wang in #168
[fix] resolve mamba init command not found in docker by @Fisher-Wang in #171
[fix] fix docker library conflict by @Fisher-Wang in #174
[update] align direct_rl_env with IsaacLab v1.4.1 version by @Fisher-Wang in #153
Update 0_static_scene.md to use sapien3 by @Fisher-Wang in #175
[fix] fix genesis non-headless usage in docker by @Fisher-Wang in #176
[fix] docker conda init by @Fisher-Wang in #177
[fix] isaacgym humanoid training pipeline by @WeisonWEileen in #164
[fix] isaacgym dict access and update docs by @WeisonWEileen in #183
[doc] support matrix by @Fisher-Wang in #182
[pre-commit.ci] pre-commit autoupdate by @pre-commit-ci[bot] in #188
RoboVerse Benchmark Update by @WinterMelooooo in #191
[fix] camera state not converted from tensor state to nested state, introduced in #159 by @Fisher-Wang in #194
[refactor] mujoco simplify code by @Fisher-Wang in #195
[add] mjx autodownload by @Fisher-Wang in #196
[pre-commit.ci] pre-commit autoupdate by @pre-commit-ci[bot] in #193
[refactor] reorganize and add docstring for base_robot_cfg by @Fisher-Wang in #197
[refactor] supported simulator by @Fisher-Wang in #198
[refactor] reorganize objects.py by @Fisher-Wang in #200
[feat] add viewer for MJX handler&fix bugs for set_states by @MurphyZhao04 in #202
[delete] unused data and doc by @Fisher-Wang in #206
[bugfix] isaacgym set joint stiffness=800, damping=40 when not specified by @Fisher-Wang in #207
[fix] mujoco&mjx cvel bugs by @MurphyZhao04 in #204
[pre-commit.ci] pre-commit autoupdate by @pre-commit-ci[bot] in #209
[update] arnold tapwater scene path by @Fisher-Wang in #210
[feat] add unitree dex3-1 by @WeisonWEileen in #190
[feature] isaacgym can read from mjcf file by @Fisher-Wang in #211
[add] shadow hand by @Fisher-Wang in #212
[fix] issue #98 on mujoco state inconsistency by @MurphyZhao04 in #215
[pre-commit.ci] pre-commit autoupdate by @pre-commit-ci[bot] in #222
[feature] mujoco support parallel envs by @Fisher-Wang in #229
[fix/isaacgym] reset pipeline by @WeisonWEileen in #216
[fix] isaacgym dt definition correct by @Fisher-Wang in #232
Feature/set timestep and use JIT forward in set_state by @MurphyZhao04 in #219
[fix] specify device in BaseChecker by @WeisonWEileen in #233
[feature] IsaacLab support multiple robots by @Fisher-Wang in #208
[performance] add caching for get_states method by @Fisher-Wang in #240
[pre-commit.ci] pre-commit autoupdate by @pre-commit-ci[bot] in https://...