| Name | objectrl JSON |
| Version |
0.1.1.post1
JSON |
| download |
| home_page | None |
| Summary | ObjectRL: An Object-Oriented Reinforcement Learning Codebase |
| upload_time | 2025-07-09 11:03:28 |
| maintainer | None |
| docs_url | None |
| author | None |
| requires_python | >=3.12 |
| license | GPL-3.0-or-later |
| keywords |
|
| VCS |
|
| bugtrack_url |
|
| requirements |
No requirements were recorded.
|
| Travis-CI |
No Travis.
|
| coveralls test coverage |
No coveralls.
|
# ObjectRL
[](https://objectrl.readthedocs.io/en/latest/)
[](https://github.com/adinlab/objectrl/blob/master/LICENSE)
<a href="https://pytorch.org/get-started/locally/"><img alt="PyTorch" src="https://img.shields.io/badge/PyTorch-ee4c2c?logo=pytorch&logoColor=white"></a>
<p align="center">
<img src="docs/_static/imgs/logo.svg" alt="ObjectRL Logo" height="150">
</p>
**ObjectRL** is a deep reinforcement learning library designed for research and rapid prototyping. It focuses on deep actor-critic algorithms for continuous control tasks such as those in the MuJoCo environment suite, while providing a flexible object-oriented architecture that supports future extensions to value-based and discrete-action methods.
---
## Features
- Object-oriented design for easy experimentation
- Implements popular deep RL algorithms for continuous control
- Includes experimental implementations of Bayesian and value-based methods
- Supports easy configuration via CLI and YAML files
- Rich examples and tutorials for customization and advanced use cases
---
## Supported Algorithms
- **DDPG** (Deep Deterministic Policy Gradient)
- **TD3** (Twin Delayed DDPG)
- **SAC** (Soft Actor-Critic)
- **PPO** (Proximal Policy Optimization)
- **REDQ** (Randomized Ensemble Double Q-Learning)
- **DRND** (Distributional Random Network Distillation)
- **OAC** (Optimistic Actor-Critic)
- **PBAC** (PAC-Bayesian Actor-Critic)
- **BNN-SAC** (Bayesian Neural Network SAC) — experimental, in examples
- **DQN** (Deep Q-Network) — experimental, in examples
---
## Installation
### Create Environment
```bash
conda create -n objectrl python=3.12 -y
conda activate objectrl
```
### Using PyPI (Recommended)
```bash
pip install objectrl
```
### From Source (Latest Development Version)
```bash
git clone https://github.com/adinlab/objectrl.git
cd objectrl
pip install -e .
```
### Optional Dependencies
To enable additional features such as documentation generation:
```bash
pip install objectrl['docs']
```
---
## Quick Start Guide
Run your first experiment using Soft Actor-Critic (SAC) on the default `cheetah` environment:
```bash
python objectrl/main.py --model.name sac
```
### Change Algorithm and Environment
Run DDPG on the `hopper` environment:
```bash
python objectrl/main.py --model.name ddpg --env.name hopper
```
### Customize Training Parameters
Train SAC for 100,000 steps and evaluate every 5 episodes:
```bash
python objectrl/main.py --model.name sac --env.name hopper --training.max_steps 100000 --training.eval_episodes 5
```
### Use YAML Configuration Files
For more complex or reproducible setups, create YAML config files in `objectrl/config/model_yamls/` and specify them at runtime:
```bash
python objectrl/main.py --config objectrl/config/model_yamls/ppo.yaml
```
Example `ppo.yaml`:
```yaml
model:
name: ppo
training:
warmup_steps: 0
learn_frequency: 2048
batch_size: 64
n_epochs: 10
```
---
## Documentation
Explore detailed documentation, tutorials, and API references at: [https://objectrl.readthedocs.io](https://objectrl.readthedocs.io)
---
## Citation
If you use ObjectRL in your research, please cite:
```bibtex
@article{baykal2025objectrl,
title={ObjectRL: An Object-Oriented Reinforcement Learning Codebase},
author={Baykal, Gulcin and Akg{\"u}l, Abdullah and Haussmann, Manuel and Tasdighi, Bahareh and Werge, Nicklas and Wu Yi-Shan and Kandemir, Melih},
year={2025},
journal={arXiv preprint arXiv:2507.03487}
}
```
Raw data
{
"_id": null,
"home_page": null,
"name": "objectrl",
"maintainer": null,
"docs_url": null,
"requires_python": ">=3.12",
"maintainer_email": null,
"keywords": null,
"author": null,
"author_email": "Abdullah Akgul <akgul@imada.sdu.dk>, Bahareh Tasdighi <tasdighi@imada.sdu.dk>, Gulcin Baykal Can <baykalg@imada.sdu.dk>, Manuel Haussmann <haussmann@imada.sdu.dk>, Melih Kandemir <kandemir@imada.sdu.dk>, Nicklas Werge <werge@sdu.dk>, Yi-Shan Wu <yswu@imada.sdu.dk>",
"download_url": "https://files.pythonhosted.org/packages/7b/91/9e0de0de2b770d4835a658e4bf0c68e40852ba7c17e8a56c7c0490a95d17/objectrl-0.1.1.post1.tar.gz",
"platform": null,
"description": "# ObjectRL\n\n[](https://objectrl.readthedocs.io/en/latest/)\n[](https://github.com/adinlab/objectrl/blob/master/LICENSE)\n<a href=\"https://pytorch.org/get-started/locally/\"><img alt=\"PyTorch\" src=\"https://img.shields.io/badge/PyTorch-ee4c2c?logo=pytorch&logoColor=white\"></a>\n\n<p align=\"center\">\n <img src=\"docs/_static/imgs/logo.svg\" alt=\"ObjectRL Logo\" height=\"150\">\n</p>\n\n**ObjectRL** is a deep reinforcement learning library designed for research and rapid prototyping. It focuses on deep actor-critic algorithms for continuous control tasks such as those in the MuJoCo environment suite, while providing a flexible object-oriented architecture that supports future extensions to value-based and discrete-action methods.\n\n---\n\n## Features\n\n- Object-oriented design for easy experimentation \n- Implements popular deep RL algorithms for continuous control \n- Includes experimental implementations of Bayesian and value-based methods \n- Supports easy configuration via CLI and YAML files \n- Rich examples and tutorials for customization and advanced use cases \n\n---\n\n## Supported Algorithms\n\n- **DDPG** (Deep Deterministic Policy Gradient) \n- **TD3** (Twin Delayed DDPG) \n- **SAC** (Soft Actor-Critic) \n- **PPO** (Proximal Policy Optimization) \n- **REDQ** (Randomized Ensemble Double Q-Learning) \n- **DRND** (Distributional Random Network Distillation) \n- **OAC** (Optimistic Actor-Critic) \n- **PBAC** (PAC-Bayesian Actor-Critic) \n- **BNN-SAC** (Bayesian Neural Network SAC) \u2014 experimental, in examples \n- **DQN** (Deep Q-Network) \u2014 experimental, in examples \n\n---\n\n## Installation\n\n### Create Environment\n\n```bash\nconda create -n objectrl python=3.12 -y\nconda activate objectrl\n```\n\n### Using PyPI (Recommended)\n\n```bash\npip install objectrl\n```\n\n### From Source (Latest Development Version)\n\n```bash\ngit clone https://github.com/adinlab/objectrl.git\ncd objectrl\npip install -e .\n```\n\n### Optional Dependencies\n\nTo enable additional features such as documentation generation:\n```bash\npip install objectrl['docs']\n```\n\n---\n\n## Quick Start Guide\n\nRun your first experiment using Soft Actor-Critic (SAC) on the default `cheetah` environment:\n\n```bash\npython objectrl/main.py --model.name sac\n```\n\n### Change Algorithm and Environment\n\nRun DDPG on the `hopper` environment:\n\n```bash\npython objectrl/main.py --model.name ddpg --env.name hopper\n```\n\n### Customize Training Parameters\n\nTrain SAC for 100,000 steps and evaluate every 5 episodes:\n\n```bash\npython objectrl/main.py --model.name sac --env.name hopper --training.max_steps 100000 --training.eval_episodes 5\n```\n\n### Use YAML Configuration Files\n\nFor more complex or reproducible setups, create YAML config files in `objectrl/config/model_yamls/` and specify them at runtime:\n\n```bash\npython objectrl/main.py --config objectrl/config/model_yamls/ppo.yaml\n```\n\nExample `ppo.yaml`:\n\n```yaml\nmodel:\n name: ppo\ntraining:\n warmup_steps: 0\n learn_frequency: 2048\n batch_size: 64\n n_epochs: 10\n```\n\n---\n\n## Documentation\n\nExplore detailed documentation, tutorials, and API references at: [https://objectrl.readthedocs.io](https://objectrl.readthedocs.io)\n\n---\n\n## Citation\n\nIf you use ObjectRL in your research, please cite:\n\n```bibtex\n@article{baykal2025objectrl,\n title={ObjectRL: An Object-Oriented Reinforcement Learning Codebase}, \n author={Baykal, Gulcin and Akg{\\\"u}l, Abdullah and Haussmann, Manuel and Tasdighi, Bahareh and Werge, Nicklas and Wu Yi-Shan and Kandemir, Melih},\n year={2025},\n journal={arXiv preprint arXiv:2507.03487}\n}\n```\n\n\n\n\n",
"bugtrack_url": null,
"license": "GPL-3.0-or-later",
"summary": "ObjectRL: An Object-Oriented Reinforcement Learning Codebase",
"version": "0.1.1.post1",
"project_urls": null,
"split_keywords": [],
"urls": [
{
"comment_text": null,
"digests": {
"blake2b_256": "c414d988b5d1e18995acf83e95042aeef494e2649087dac0e2f07a22ed2aa538",
"md5": "dd3197319ce0158170fedd6a784d959c",
"sha256": "c23db6a212a9c0d9a625ebda46908b2d36798433d28a6c9db760d253c014fff9"
},
"downloads": -1,
"filename": "objectrl-0.1.1.post1-py3-none-any.whl",
"has_sig": false,
"md5_digest": "dd3197319ce0158170fedd6a784d959c",
"packagetype": "bdist_wheel",
"python_version": "py3",
"requires_python": ">=3.12",
"size": 125301,
"upload_time": "2025-07-09T11:03:27",
"upload_time_iso_8601": "2025-07-09T11:03:27.120743Z",
"url": "https://files.pythonhosted.org/packages/c4/14/d988b5d1e18995acf83e95042aeef494e2649087dac0e2f07a22ed2aa538/objectrl-0.1.1.post1-py3-none-any.whl",
"yanked": false,
"yanked_reason": null
},
{
"comment_text": null,
"digests": {
"blake2b_256": "7b919e0de0de2b770d4835a658e4bf0c68e40852ba7c17e8a56c7c0490a95d17",
"md5": "0521521f84ae178fc4884ba655b9e639",
"sha256": "a27b9d010d105b72925583eaf29c07e29b67499ae132af8be38ee55fc0340670"
},
"downloads": -1,
"filename": "objectrl-0.1.1.post1.tar.gz",
"has_sig": false,
"md5_digest": "0521521f84ae178fc4884ba655b9e639",
"packagetype": "sdist",
"python_version": "source",
"requires_python": ">=3.12",
"size": 79317,
"upload_time": "2025-07-09T11:03:28",
"upload_time_iso_8601": "2025-07-09T11:03:28.938620Z",
"url": "https://files.pythonhosted.org/packages/7b/91/9e0de0de2b770d4835a658e4bf0c68e40852ba7c17e8a56c7c0490a95d17/objectrl-0.1.1.post1.tar.gz",
"yanked": false,
"yanked_reason": null
}
],
"upload_time": "2025-07-09 11:03:28",
"github": false,
"gitlab": false,
"bitbucket": false,
"codeberg": false,
"lcname": "objectrl"
}
