All notable changes to this project will be documented in this file.
The format is based on Keep a Changelog, and this project adheres to Semantic Versioning.
-
Feat(off-policy): support off-policy pid and update performance for navigation by @Jiayi Zhou in PR #245.
-
Style(model-based): fix mypy and polish api docstring by @Jiayi Zhou in PR #244.
-
Feat: improve test coverage and clear redundant code by @Jiayi Zhou in PR #238.
-
Feat: update benchmarks and provide configs for reproducing results by @Jiayi Zhou in PR #238.
-
Feat: add CODEOWNERS and refine ISSUE TEMPLATE by @Jiaming Ji in PR #233.
-
Style: support mypy checking and update docstring style by @Jiayi Zhou in PR #221.
-
Feat: add Dockerfile and codecov.yml by @XuehaiPan in PR #217.
-
Chore: update benchmark performance for first-order algorithms by @Borong Zhang in PR #215.
-
Chore: clean some trivial code by @Borong Zhang in PR #214.
-
Feat(pid-lagrange, test): add algo and update test by @Jiayi Zhou in PR #210.
-
Feat: add offline algorithms by @Ruiyang Sun in PR #175.
-
Docs: polish algorithms tutorial by @Jiayi Zhou in PR #242.
-
Docs: change link to PKU-Alignment by @Jiayi Zhou in PR #239.
-
Docs: polish readme by @Jiaming Ji in PR #231.
-
Docs: polish algorithm tutorial and update API docs by @Jiayi Zhou in PR #225.
- Fix: fix adapter device and exp grid by @Jiayi Zhou in PR #243.
-
Style: fix grammar in README and normalize string in pyproject.toml by @XuehaiPan.
-
Feat: add model-based algorithms by @Weidong Huang in PR #212.
-
Feat(saute, simmer): support saute rl and clean the code by @Jiayi Zhou in PR #209.
-
Feat(off-policy): support off-policy lag by @Jiayi Zhou in PR #204.
-
Chore: upload tutorial by @Borong Zhang in PR #201.
-
Deps(joblib): unpin joblib by @Borong Zhang in PR #197.
-
Fix(simmer, PID): fix simmer performance and tune parameters by @Jiayi Zhou in PR #220.
-
Fix: enable smooth param in Costs when plotting by @Borong Zhang in PR #208.
-
Fix(off-policy): fix log when not update by @Jiayi Zhou in PR #206.
-
Fix: check duplicated parameters and values which are specified in experiment grid by @Borong Zhang in PR #203.
-
Docs: update contributing.md by @Jiaming Ji in PR #227.
-
Docs: fix small typo in README.md by @mickelliu in PR #211.
-
Docs: change link to PKU-Alignment by @Jiaming Ji in PR #205.
-
Refactor: minor changes regarding documentation and dependencies by @Jiaming Ji in PR #219.
-
Refactor(common/logger): refactor and simplify logger storage logic by @XuehaiPan in PR #216.
-
Refactor(algo_wrapper, configs): rename update cycle and refactor structure by @Jiayi Zhou in PR #213.
-
Refactor: update hyper-parameters for first-order algorithms by @Borong Zhang in PR #199.
-
Refactor: condense top-level benchmarks by @Jiaming Ji in PR #198.
-
Feat: update CLI for gpu and statistics tools by @Borong Zhang in PR #192.
-
Feat: add
ruff
andcodespell
integration by @XuehaiPan in PR #186. -
Feat: perfecting training workflow by @Borong Zhang in PR #185.
-
Deps(torch): pin the version of pytorch by @Borong Zhang in PR #184.
- Fix(experiment grid): fix file path problem when using gpu in experiment grid by @Borong Zhang in PR #194.
-
Docs: update api documents by @Jiayi Zhou in PR #191.
-
Docs: update changelog.md by @Jiaming Ji in PR #187.
- Add MANIFEST.in by @Borong Zhang in PR #182.
- Update api documentation by @Jiayi Zhou in PR #181
- Feat(statistics tools): support statistics tools for experiments by @Borong Zhang in PR #157.
- Support cuda by @Jiayi Zhou in PR #163.
- Support command line interfaces for OmniSafe by @Borong Zhang in PR #144.
- Refactor(wrapper): refactor the cuda setting by @Jiayi Zhou in PR #176.
- Fix(onpolicy_adapter): fix the calculation of last state value by @Borong Zhang in PR #164.
- Fix(config.py): fix config assertion by @Jiayi Zhou in PR #174.
- Fix autoreset wrapper in by @r-y1 PR #167.
- Update docs style by @Jiayi Zhou in PR #169.
- Fix typo in readme by @Ruiyang Sun in PR #172.
- Update README and the usage of CLI by @Jiaming Ji in PR #138.
- Check out
Development
for more.
- Chore(on-policy): update benchmark performance for first-order algorithms by @Borong Zhang in PR #148.
- Feat(off-policy): add DDPG, TD3 SAC by @Jiayi Zhou in PR #128.
- Feat: support policy evaluation by @Jiayi Zhou in PR #137.
- Test: add more test case, and fix bugs by @Ruiyang Sun in PR #136.
- Fix(logger, wrapper): support csv file and velocity tasks by @Jiayi Zhou in PR #131.
- Feat: update architecture of config.yaml by @Jiaming Ji in PR #126.
- Chore: support num_thread setting by @Jiayi Zhou in PR #124.
- Refactor: change architecture of omnisafe by @Ruiyang Sun in PR #121.
- Fix(on-policy): fix the second order algorithms performance by @Jiayi Zhou in PR #147.
- Fix(rollout, exp_grid): fix logdir path conflict by @Borong Zhang in PR #145.
- Fix: support new config for exp_grid by @Borong Zhang in PR #142.
- Fix(ppo): fix entropy loss by @Jiayi Zhou in PR #135.
- Fix(algo): fix no return in algo_wrapper::learn by @Ruiyang Sun in PR #122.
- Docs: Update changelog by @Jiaming Ji.
- Docs: Update README.md: fix action passing by @Jiayi Zhou in PR #149.
- Chore: fix typo by @1Asan in PR #134.
- Fix(P3O): fix P3O performance by @Jiayi Zhou in PR #123.
- Fix(off-policy): fix
action passing
by @Jiayi Zhou in PR #119.
- Docs: update logo by @Jiayi Zhou in PR #125.
- Fix(evaluator): fix evaluator by @Ruiyang Sun in PR #117.
- Build(env): delete local
safety-gymnaisum
dependence by @Ruiyang Sun in PR #102. - Refactor(buffer): refactor
buffer
by @Ruiyang Sun in PR #101. - Refactor: clean the code by @Jiayi Zhou in
- Fix: fix tools by @Jiayi Zhou in PR #100.
- Fix: fix algo wrapper by @Jiayi Zhou in PR #99. PR #97.
- Modify
logo.png
and addrequirements.txt
by @Ruiyang Sun in PR #103.
- Chore: update linter settings by @XuehaiPan.
- Chore: update ci by @Jiayi Zhou in PR #90 reviewed by @Jiaming Ji and @friedmainfunction.
- Chore: update yaml by @Jiayi Zhou in PR #92 and #93 reviewed by @Jiaming Ji and @friedmainfunction.
- Refactor(objects): change object type into free_geom by @Borong Zhang in PR #89.
- Chore: update algorithms configuration by @Jiayi Zhou in PR #88.
- Feat: support cuda by [@Jiayi Zhou](https://github.com/Gaiejj in PR #86.
- Feat(render): add keyboard debug mode for some agents in all tasks by @Borong Zhang in PR #83.
- Feat: add experiment grid by @Jiaming Ji in PR #84.
- Fix seed setting by @Jiayi Zhou in PR #82.
- Feat(agents): add
ant
agent by @Borong Zhang in PR #82. - Refactor(safety-gymnaisum):
code decoupling
by @Borong Zhang in PR #81. - Feat: add new algorithm by @Jiayi Zhou in PR #80.
- Refactor: change wrapper setting by @Jiayi Zhou in PR #73.
- Feat:
vectorized
environment by @Jiayi Zhou in PR #74.
- Feat(agents, tasks, Evaluator): support
circle012
and new agentracecar
, update evaluator by @Borong Zhang in PR #59.
- Refactor: enhanced model-based code, add
CAP
algorithm by @Weidong Huang in PR #59. - Feat: support auto render as .mp4 videos, add examples and tests by @Borong Zhang in PR #60.
- Fix(model-based): fix cap cost bug and lag beta value in cap.yaml by @Weidong Huang in PR #62.
- Fix(render): fix markers are not shown in the rgb array returned by env.render() by @Borong Zhang in PR #61.
- Feat(circle, run): support new tasks by @Borong Zhang in PR #50.
- Add Makefile by @XuehaiPan in PR #53.
- Fix bug for namedtuple by @Jiayi Zhou in PR #54.
- Fix spelling error by @Jiayi Zhou in PR #56, reviewed by @Jiaming Ji and @XuehaiPan.
- Refactor: open pylint in pre-commit by @Jiaming Ji in PR #48.
- Refactor: change the details and yaml files of on policy algorithm by @Jiayi Zhou in PR #41.
- Feat: add CUP algorithm by @Jiayi Zhou in PR #43.
- Feat(wrapper): separated wrapper for different algorithmic environments by @Jiaming Ji in PR #44.
- Chore: rename files and enable pylint by @Borong Zhang in PR #39.
- Retouch the formatting and add PPO docs for OmniSafe by @Jiayi Zhou in PR #40.
- Add Lagrangian method documentation by @Jiayi Zhou in PR #42.
- Refactor(README): show the implemented algorithms in more detail by @Jiaming Ji in PR #47.
- Refactor: more OOP style code were used and made better code and file structure by @Borong Zhang in PR #37.
- Refactor: change the file layout of omnisafe by @Jiaming Ji in PR #35.
- Fix(env_wrapper): fix warning caused by 'none' string default value by @Borong Zhang in PR #30.
- Docs: retouch the formatting and add links to the formula numbers by @Jiayi Zhou in PR #31.
- Chore(.github): update issue templates by @XuehaiPan in PR #29.
- Refactor packaging by @XuehaiPan in PR #20.
- Add ddpg, clean some code, modify algo_wrapper in PR #24 by @Jiaming Ji.
- Add
TRPO
to docs by @Jiayi Zhou in PR #28. - Add
FOCOPS
andPCPO
to docs by @XuehaiPan in #21.
- Add render_mode:
human
,rgb_array
,depth_array
in safety-gymnasium:safety_gym_v2
. - Add Model-based Safe Algorithms:
mbppolag
,safeloop
by @Weidong Huang in #12. - Add .editorconfig and update license by @XuehaiPan in #8.
- Fix readme typo by @erjanmx in PR #13.
- Fix ambiguous config yaml for algorithms by @Jiaming Ji in PR #6.
- Fix vis
safety_gym_v2
with del the render_mode by @Jiaming Ji in PR #3.
- Refactor some code in omnisafe, add
CHANGELOG.md
,` and del install.md and tutorial in PR #16 by @Jiaming Ji. - Docs: add
PCPO
in omnisafe's docs and modifyCPO
by @Jiayi Zhou in #9. - Add
CPO
andIntro
in omnisafe's docs by @Jiayi Zhou in PR #7. - Add render mode and vision input in safety-gymnasium:
safety_gym_v2
by @Jiaming Ji in PR #5.