v0.5.0

Features

• Feat(off-policy): support off-policy pid and update performance for navigation by @Gaiejj in PR #245 .
• Style(model-based): fix mypy and polish api docstring by @Gaiejj in PR #244.
• Feat: improve test coverage and clear redundant code by @Gaiejj in PR #238.
• Feat: update benchmarks and provide configs for reproducing results by @Gaiejj in PR #238
• Feat: add CODEOWNERS and refine ISSUE TEMPLATE by @zmsn-2077 in PR #233.
• Style: support mypy checking and update docstring style by @Gaiejj in PR #221.
• Feat: add Dockerfile and codecov.yml by @XuehaiPan in PR #217.
• Chore: update benchmark performance for first-order algorithms by @muchvo in PR #215.
• Chore: clean some trivial code by @muchvo in PR #214.
• Feat(pid-lagrange, test): add algo and update test by @Gaiejj in PR #210.
• Feat: add offline algorithms by @rockmagma02 in PR #175.

Documentations

• Docs: polish algorithms tutorial by @Gaiejj in PR #242.
• Docs: change link to PKU-Alignment by @Gaiejj in PR #239.
• Docs: polish readme by @zmsn-2077 in PR #231.
• Docs: polish algorithm tutorial and update API docs by @Gaiejj in PR #225.

Fixes

• Fix: fix adapter device and exp grid by @Gaiejj in PR #243.

v0.4.0

1. refactor: condense top-level benchmarks
2. deps(joblib): unpin joblib
3. fix: check duplicated parameters and values which are specified in the experiment grid
4. refactor: update hyper-parameters for first-order algorithms
5. feat(off-policy): support off-policy lag
6. fix(off-policy): fix log when not update
7. fix: enable smooth param in Costs when plotting
8. feat(saute, simmer): support saute rl and clean the code
9. refactor(algo_wrapper, configs): rename update cycle and refactor structure
10. style: fix grammar in README and normalize string in pyproject.toml
11. refactor(common/logger): refactor and simplify logger storage logic
12. chore: update benchmark performance for first-order algorithms
13. fix(simmer, PID): fix simmer performance and tune parameters
14. refactor: minor changes regarding documentation and dependencies
15. feat: add model-based algorithms
16. docs: update contributing.md

v0.3.0

1. docs: update api documents
2. fix(experiment grid): fix file path problem when using gpu in experiment grid
3. feat: update CLI for gpu and statistics tools
4. feat: perfecting training workflow
5. deps(torch): pin the version of pytorch
6. docs: update changelog.md
7. feat: add ruff and codespell integration

v0.2.2

1. chore: add MANIFEST.in
2. docs: update API documentation

v0.2.1

1. support statistics tools for experiments launched by omnisafe
2. fix some typos

v0.2.0

1. chore: update GitHub workflow
2. feat: support command line interfaces for omnisafe
3. fix(onpolicy_adapter): fix the calculation of last state value
4. feat: support cuda
5. docs: update docs style
6. fix(config.py): fix config assertion
7. fix: fix autoreset wrapper
8. refactor(wrapper): refactor the cuda setting
9. docs: update README and the usage of CLI

v0.1.0

The first stable release of omnisafe.