https://github.com/facebookresearch/rlmeta

RLMeta is a light-weight flexible framework for Distributed Reinforcement Learning Research.
https://github.com/facebookresearch/rlmeta

Last synced: 3 months ago
JSON representation

RLMeta is a light-weight flexible framework for Distributed Reinforcement Learning Research.

• Host: GitHub
• URL: https://github.com/facebookresearch/rlmeta
• Owner: facebookresearch
• License: mit
• Archived: true
• Created: 2021-12-16T02:47:46.000Z (over 2 years ago)
• Default Branch: main
• Last Pushed: 2023-02-11T06:20:06.000Z (over 1 year ago)
• Last Synced: 2024-03-04T15:33:26.825Z (4 months ago)
• Language: Python
• Size: 381 KB
• Stars: 285
• Watchers: 14
• Forks: 28
• Open Issues: 9
• Metadata Files:
  • Readme: README.md
  • Contributing: CONTRIBUTING.md
  • License: LICENSE
  • Code of conduct: CODE_OF_CONDUCT.md
  • Citation: CITATION.cff

Lists

• awesome-production-machine-learning - RLMeta - RLMeta is a flexible lightweight research framework for Distributed Reinforcement Learning based on PyTorch and moolib (Industry Strength RL)

README

        
# RLMeta

rlmeta - a flexible lightweight research framework for Distributed 

Reinforcement Learning based on [`PyTorch`](https://pytorch.org/) and 

[`moolib`](https://github.com/facebookresearch/moolib)

## Installation

To build from source, please install [`PyTorch`](https://pytorch.org/) first, 

and then run the commands below.

```

$ git clone https://github.com/facebookresearch/rlmeta

$ cd rlmeta

$ git submodule sync && git submodule update --init --recursive

$ pip install -e .

```

## Run an Example

To run the example for Atari Pong game with PPO algorithm:

```

$ cd examples/atari/ppo

$ python atari_ppo.py env.game="Pong" num_epochs=20

```

We are using [`hydra`](https://hydra.cc/) to define configs for trainining jobs.

The configs are defined in

```

./conf/conf_ppo.yaml

```

The logs and checkpoints will be automatically saved to

```

./outputs/{YYYY-mm-dd}/{HH:MM:SS}/

```

After training, we can draw the training curve by run

```

$ python ../../plot.py --log_file=./outputs/{YYYY-mm-dd}/{HH:MM:SS}/atari_ppo.log --fig_file=./atari_ppo.png --xkey=time

```

One example of the training curve is shown below.

![atari_ppo](./docs/source/_static/img/atari_ppo.png)

## License

rlmeta is licensed under the MIT License. See [`LICENSE`](LICENSE) for details.