Pref-RL

KO | EN

gitlite — search

#python #hacktoberfest #ai #typescript #go #deep-learning #linux #docker #macos #kubernetes #vue #mcp #openai #open-source

(★ 11)

Pref-RL provides ready-to-use PbRL agents that are easily extensible.

Pref-RL 최신버젼 다운로드

최종 버전 다운로드 (.zip)

// repository documentation

Was this content helpful?

★ 0 (0 ratings)

# Pref-RL
`Pref-RL` provides ready-to-use PbRL agents that are easily extensible.

We strive for:
- Training of state-of-the-art PbRL agents on arbitrary environments in a few lines of code.
- An easily extensible agent framework to quickly build your own custom agents on top.
- A clean and well-maintained implementation (in Python).

## Main features (planned)
**Note:** The project is still in an experimental development phase. 
The initial feature set is not yet completed and no performance tests have been conducted.

#### General
- [x] Simple training of deep PbRL agents on arbitrary Gym environments 
- [x] FNN and CNN reward models (implemented in [PyTorch](https://pytorch.org/))
- [x] Synthetic preference data generation
- [ ] Human preference data generation / collection (under development)
- [x] State-of-the-art RL algorithms (via [Stable Baselines3](https://github.com/DLR-RM/stable-baselines3))
- [x] TensorBoard support

#### Bould your own agents
- [x] Custom environments ([Open AI Gym](https://gym.openai.com/) compatible) 
- [x] Custom reward models 
- [x] Custom PbRL agents with almost no code
- [x] Easy integration of custom components

#### Code quality
- [x] High code coverage (> 90%)
- [ ] PEP8 code style
- [ ] Type hints 
- [ ] Learning performance benchmarked against state-of-the-art

#### Other features
- [ ] Active, ensemble-based query selection
- [ ] Advanced reward model pretraining with IRL, intrinsic motivation, ...
- [ ] [PEBBLE](https://github.com/pokaxpoka/B_Pref) PbRL algorithm

## Installation
These instructions presume a *nix or OS X operating system.

### Prerequisites
This framework requires Python 3.6+ and `pip`.

Install [`pip`](http://www.pip-installer.org/en/latest/) with these
[installation instructions](http://www.pip-installer.org/en/latest/installing.html).

### Install using pip
<a id="install-with-pip"></a>
Install the requirements using pip:
```
pip install -r requirements.txt
```
### Windows
On Windows, you may encounter issues running OpenAI Gym Atari environments.
[This stack overflow answer](https://stackoverflow.com/a/46739299/3902240)
could help.

## Example
See teach.py for an example of how to instantiate and run an agent. 
When you're at the project's root, use the following to run a default agent from the command line:
```
python teach.py --env_id "CartPole-v1" --reward_model "Mlp" --num_rl_timesteps 200000 --num_pretraining_preferences 100
```

You can monitor agent training with TensorBoard. Start it with:
```
tensorboard --logdir=runs
```
View the output by navigating to https://localhost:6006.

## Testing the implementation
All unit tests in the framework can be run using `pytest`. 
It is part of the project's requirements and has therefore already been installed. 
Run the tests with:
```
pytest ./tests/
```

Pref-RL

Pref-RL 최신버젼 다운로드

오늘의 추천 프로젝트

Was this content helpful?

Recent Feedback

Pref-RL

Pref-RL 최신버젼 다운로드

오늘의 추천 프로젝트

Was this content helpful?

Recent Feedback

Download README