Ecosyste.ms: Awesome
An open API service indexing awesome lists of open source software.
https://github.com/zuoxingdong/lagom
lagom: A PyTorch infrastructure for rapid prototyping of reinforcement learning algorithms.
https://github.com/zuoxingdong/lagom
artificial-intelligence cem cmaes ddpg deep-deterministic-policy-gradient deep-learning deep-reinforcement-learning evolution-strategies machine-learning mujoco policy-gradient ppo proximal-policy-optimization python pytorch reinforcement-learning research sac soft-actor-critic td3
Last synced: 2 months ago
JSON representation
lagom: A PyTorch infrastructure for rapid prototyping of reinforcement learning algorithms.
- Host: GitHub
- URL: https://github.com/zuoxingdong/lagom
- Owner: zuoxingdong
- License: mit
- Created: 2017-12-21T09:51:28.000Z (about 7 years ago)
- Default Branch: master
- Last Pushed: 2022-11-19T05:52:37.000Z (about 2 years ago)
- Last Synced: 2024-08-04T00:13:12.757Z (6 months ago)
- Topics: artificial-intelligence, cem, cmaes, ddpg, deep-deterministic-policy-gradient, deep-learning, deep-reinforcement-learning, evolution-strategies, machine-learning, mujoco, policy-gradient, ppo, proximal-policy-optimization, python, pytorch, reinforcement-learning, research, sac, soft-actor-critic, td3
- Language: Jupyter Notebook
- Homepage:
- Size: 95.9 MB
- Stars: 373
- Watchers: 16
- Forks: 30
- Open Issues: 16
-
Metadata Files:
- Readme: README.md
- License: LICENSE
Awesome Lists containing this project
- Awesome-pytorch-list-CNVersion - lagom
- Awesome-pytorch-list - lagom
README
lagom
A PyTorch infrastructure for rapid prototyping of reinforcement learning algorithms.
[lagom](https://sv.wikipedia.org/wiki/Lagom) is a 'magic' word in Swedish, *inte för mycket och inte för lite, enkelhet är bäst* (*not too much and not too little, simplicity is often the best*). It is the philosophy on which this library was designed.
## Why to use lagom ?
`lagom` balances between the flexibility and the usability when developing reinforcement learning (RL) algorithms. The library is built on top of [PyTorch](https://pytorch.org/) and provides modular tools to quickly prototype RL algorithms. However, it does not go overboard, because too low level is often time consuming and prone to potential bugs, while too high level degrades the flexibility which makes it difficult to try out some crazy ideas fast.
We are continuously making `lagom` more 'self-contained' to set up and run experiments quickly. It internally supports base classes for multiprocessing ([master-worker framework](https://en.wikipedia.org/wiki/Master/slave_(technology))) for parallelization (e.g. experiments and evolution strategies). It also supports hyperparameter search by defining configurations either as grid search or random search.
**Table of Contents**
- [Installation](#installation)
- [Install dependencies](#install-dependencies)
- [Install lagom from source](#install-lagom-from-source)
- [Documentation](#documentation)
- [RL Baselines](#rl-baselines)
- [How to use lagom](#how-to-use-lagom)
- [Examples](#examples)
- [Test](#test)
- [What's new](#What's-new)
- [Reference](#reference)
## Installation
We highly recommand using an Miniconda environment:
```bash
conda create -n lagom python=3.7
```
### Install dependencies
```bash
pip install -r requirements.txt
```
We also provide some bash scripts in [scripts/](scripts/) directory to automatically set up the system configurations, conda environment and dependencies.
### Install lagom from source
```bash
git clone https://github.com/zuoxingdong/lagom.git
cd lagom
pip install -e .
```
Installing from source allows to flexibly modify and adapt the code as you pleased, this is very convenient for research purpose.
## Documentation
The documentation hosted by ReadTheDocs is available online at [http://lagom.readthedocs.io](http://lagom.readthedocs.io)
## RL Baselines
We implemented a collection of standard reinforcement learning algorithms at [baselines](baselines/) using lagom.
## How to use lagom
A common pipeline to use `lagom` can be done as following:
1. Define your [RL agent](lagom/agent.py)
2. Define your [environment](lagom/envs)
3. Define your [engine](lagom/engine.py) for training and evaluating the agent in the environment.
4. Define your [Configurations](lagom/experiment/config.py) for hyperparameter search
5. Define `run(config, seed, device)` for your experiment pipeline
6. Call `run_experiment(run, config, seeds, num_worker)` to parallelize your experiments
A graphical illustration is coming soon.
### Examples
We provide a few simple [examples](examples/).
## Test
We are using [pytest](https://docs.pytest.org) for tests. Feel free to run via
```bash
pytest test -v
```
## What's new
- 2019-03-04 (v0.0.3)
- Much easier and cleaner APIs
- 2018-11-04 (v0.0.2)
- More high-level API designs
- More unit tests
- 2018-09-20 (v0.0.1)
- Initial release
## Reference
This repo is inspired by [OpenAI Gym](https://github.com/openai/gym/), [OpenAI baselines](https://github.com/openai/baselines), [OpenAI Spinning Up](https://github.com/openai/spinningup)
Please use this bibtex if you want to cite this repository in your publications:
@misc{lagom,
author = {Zuo, Xingdong},
title = {lagom: A PyTorch infrastructure for rapid prototyping of reinforcement learning algorithms},
year = {2018},
publisher = {GitHub},
journal = {GitHub repository},
howpublished = {\url{https://github.com/zuoxingdong/lagom}},
}