OmniGibson v1.1.0
We are excited to announce the latest release of OmniGibson, bringing significant enhancements to our simulation platform for embodied AI research and development.
Highlights
- Fast Installation with pip: You can now install OmniGibson and Isaac Sim via pip! For detailed
installation instructions, please visit our official documentation. - Multiple/Parallel Environments: Support for reinforcement learning with optimized performance
by running multiple environments at once as in Isaac Gym. - Data Collection and Playback: You can now use our data collection environment wrappers for demonstration data.
- Primitive Skill Library: We now provide more robust action primitives and faster IK, with primitives
now covered under integration tests for reliability. - Torch Integration: OmniGibson now works exclusively with
torch
tensors instead ofnumpy
arrays, in
anticipation of full GPU-based simulation and data pipeline. - cuRobo Integration: OmniGibson now supports some accelerated motion planning capabilities from curobo.
This integration will be extended in the near future to fully remove the OMPL dependency in primitives. - Scene Graph Generation: OmniGibson can now generate scene graphs, compatible with single or multiple agents.
- New Robots: We support a number of new robots and provide improvements for our existing robots.
- Documentation Overhaul: Detailed OG modules and tutorials are now provided in the documentation.
Breaking Changes
- OmniGibson now uses
torch
instead ofnumpy
as is numerical computation backend, in anticipation of
full GPU-based simulation features that we plan to include in the next release. This means that users
now need to provide torch tensors when calling APIs instead of numpy arrays. - OmniGibson is no longer compatible with Isaac Sim 2023.1.1 - we recommend upgrading to 4.1. OmniGibson
is not yet compatible with Isaac Sim 4.2. - Local pose APIs have been removed and unified into the
set_position_orientation
andget_position_orientation
APIs with theframe
argument. - Absolute prim paths may no longer be provided when creating objects and prims. You will need to provide
a relative prim path which will be applied on top of the object's scene's prim, in order to support
multiple environments when necessary. - Inverse Kinematics controllers have been updated to use a local approximation ("differential IK")
instead of solving the full IK problem on every step, providing significantly better speed especially
in multi-environment contexts.
We encourage all users to upgrade to this latest version to take advantage of these improvements and new features. As always, we welcome your feedback and contributions to help make OmniGibson even better.
Merged PRs
- Replicator hot fix by @hang-yin in #667
- Optimize segmentation remapping performance by @hang-yin in #669
- Remove dev docker image, add colab image & improve example/installation usability by @cgokmen in #670
- Black formatter by @hang-yin in #672
- Add black precommit hook & use pre-commit CI for linter by @cgokmen in #673
- Check Isaac Sim version by @hang-yin in #680
- Fix tests for v1.0.0 asset release by @ChengshuLi in #696
- Update profiling by @wensi-ai in #678
- Update intrinsic matrix by @hang-yin in #692
- Fix primitives tests by @sujaygarlanka in #701
- Feat/docs update v2 by @cremebrule in #693
- Fix/general improvements by @cremebrule in #698
- Run tests faster (and make them easier to re-run) by splitting them up by @cgokmen in #682
- Make the README test badge show main instead of og-develop by @cgokmen in #675
- Remove global config by @hang-yin in #702
- Require positive scales and validate local transforms by @cgokmen in #636
- Add and apply isort by @cgokmen in #674
- lazy access of gm values by @hang-yin in #631
- remove friction override for tiago by @cremebrule in #710
- fix small seg semantic id bug by @cremebrule in #722
- Primitives merge to rl by @cgokmen in #728
- Merge experimental RL updates into Hang's PR by @cgokmen in #729
- add flexibility to bddl sampling by @cremebrule in #732
- Merge franka family by @wensi-ai in #725
- fix render_product syncing when updating camera width / height by @cremebrule in #735
- Parallel environments support & RL toolkit by @cgokmen in #699
- [WIP] RL training infra for primitives by @sujaygarlanka in #282
- Fix geodesic_reward==None bug by @hang-yin in #718
- Compatibility fixes for Isaac Sim 4.0.0 by @cgokmen in #776
- Revive transition rules by @ChengshuLi in #781
- Update Docker images to Isaac Sim 4.0.0, add both-version compatibility to kit files by @cgokmen in #778
- fix typo in controller base by @ChengshuLi in #788
- Fix gymnasium flatdim by @hang-yin in #792
- Doc update multiple envs by @hang-yin in #789
- Improve documentation for Tasks, Save/Load, BEHAVIOR Tasks and Knowledgebase by @ChengshuLi in #795
- Cem docs by @cgokmen in #798
- Add import robot and speed optimization tutorials by @wensi-ai in #793
- Some quick docs fixes by @cgokmen in #799
- Feat/doc update josiah by @cremebrule in #794
- Feat/auto robot idx by @cremebrule in #797
- Remove outdated examples README, fix geodesic potential, revive behavior task test by @ChengshuLi in #804
- Improve ControllableObjectViewAPI to work with different robot types by @cgokmen in #803
- Fix scene graph by @hang-yin in #740
- Clean up launch / clear interface and fix some bugs by @cgokmen in #809
- Support OmniGibson and Isaac Sim installation via pip by @yyf20001230 in #796
- Reorganize docs and fix some small bugs by @cgokmen in #813
- Fix draw boudning box example by @hang-yin in #816
- Fix two controller issues by @cgokmen in #817
- Documentation for remote streaming by @hang-yin in #712
- isaacsim new version support and windows installation fix by @yyf20001230 in #829
- Bbox test by @hang-yin in #821
- Add robot load and drive test by @hang-yin in #819
- Support multiple agent in scene graph by @hang-yin in #836
- Data Collection / Playback Environments by @cremebrule in #811
- fix jacobian bug by @cremebrule in #838
- fix demo collection bugt by @wensi-ai in #840
- NumPy -> Torch by @hang-yin in #780
- Disable gravity compensation by @cgokmen in #855
- Fix base footprint offset not applied in ControllableObjectViewAPI by @cgokmen in #857
- add outer loop to simulator by @cremebrule in #864
- Add differential IK option to do IK without Lula by @cgokmen in #858
- Fix/osc by @cremebrule in #863
- do not require floor plane by @cremebrule in #869
- remove lula from IK controller; always use differential IK by @ChengshuLi in #876
- fix clothification for scaled cloth by @ChengshuLi in #860
- Torch mini followup - bug fixes by @hang-yin in #867
- Unify world/local/scene pose getters and setters by @yyf20001230 in #786
- update dependency versions, remove torchvision dep by @cremebrule in #881
- New robot (R1 & A1) with robot fixes by @wensi-ai in #868
- Feat/curobo by @cremebrule in #878
- Various small fixes by @cgokmen in #889
- limit numpy <2.0 by @cremebrule in #894
- Bug fixes by @hang-yin in #885
- add sanity check for object-level scale, robot can accept scale as Iterable by @ChengshuLi in #896
- Replicator hot fix by @hang-yin in #900
- Rewrite particle modifier demo by @yyf20001230 in #779
- Minor update to isaac sim setup file by @yyf20001230 in #886
- Eric's release fixes by @ChengshuLi in #905
- Pre-release bug fixes by @hang-yin in #901
- Pre release doc update by @hang-yin in #917
- PyPI release infra, version bump by @cgokmen in #919
- Improved installation via an install script instead of setup.sh and metapackage by @cgokmen in #910
- Fix/misc bugs by @cremebrule in #920
- Revive action primitives examples and tests by @yyf20001230 in #842
- Fix docker build issue due to deprecated mamba arg by @cgokmen in #921
- Remove dependency on numba by @cgokmen in #924
- Replicator Patch for INVALID segmentation label by @hang-yin in #923
- Bump asset URL version by @cgokmen in #925
New Contributors
- @yyf20001230 made their first contribution in #796
Full Changelog: v1.0.0...v1.1.0