Metadata-Version: 2.1
Name: machin
Version: 0.3.0
Summary: Reinforcement learning library
Home-page: https://github.com/iffiX/machin
Author: Iffi
Author-email: iffi@mail.beyond-infinity.com
License: MIT
Description: <div align="center">
        	<a href="https://github.com/iffiX/machin">
        		<img width="auto" height="200px" src="https://machin.readthedocs.io/en/latest/_static/icon.svg">
        	</a>
        </div>
        
        <br/>
        
        
        <div align="center">
        	<a href="https://machin.readthedocs.io/en/latest/">
        		<img alt="Read the Docs" src="https://img.shields.io/readthedocs/machin">
        	</a>
        	<a href="http://ci.beyond-infinity.com/jenkins/blue/organizations/jenkins/machin/branches/">
        		<img alt="Jenkins build" src="https://img.shields.io/jenkins/build?jobUrl=http%3A%2F%2Fci.beyond-infinity.com%2Fjenkins%2Fjob%2Fmachin%2Fjob%2Fmaster%2F">
        	</a>
        	<a href="http://ci.beyond-infinity.com/jenkins/blue/organizations/jenkins/machin/branches/">
        		<img alt="Jenkins coverage" src="https://img.shields.io/jenkins/coverage/cobertura?jobUrl=http%3A%2F%2Fci.beyond-infinity.com%2Fjenkins%2Fjob%2Fmachin%2Fjob%2Frelease%2F">
        	</a>
        	<a href="https://github.com/iffiX/machin">
        		<img alt="License" src="https://img.shields.io/github/license/iffiX/machin">
        	</a>
        	<a href="http://ci.beyond-infinity.com/reports/machin/">
        		<img alt="Report" src="https://img.shields.io/badge/report-allure-blue">
        	</a>
        	
        	
        </div>
        
        <br/>
        
        <div align="center">
        <p><strong>Readable, Reusable, Extendable</strong></p>
        
        <p><strong>Machin</strong> is a reinforcement library designed for pytorch.</p> 
        </div>
        
        <br/>
        
        ### Supported Models
        ---
        **Anything**, including recurrent networks.
        
        ### Supported algorithms
        ---
        Currently Machin has implemented the following algorithms, the list is still growing:
        
        #### Single agent algorithms:
        * [Deep Q-Network (DQN)](https://storage.googleapis.com/deepmind-media/dqn/DQNNaturePaper.pdf)
        * [Double DQN](https://arxiv.org/pdf/1509.06461.pdf)
        * [Dueling DQN](https://arxiv.org/abs/1511.06581)
        * [RAINBOW](https://arxiv.org/abs/1710.02298)
        * [Deep Deterministic policy Gradient (DDPG)](https://arxiv.org/pdf/1509.02971.pdf)
        * [Twin Delayed DDPG (TD3)](https://arxiv.org/pdf/1802.09477.pdf)
        * [Hystereric DDPG (Modified from Hys-DQN)](https://hal.archives-ouvertes.fr/hal-00187279/document)
        * [Advantage Actor-Critic (A2C)](https://openai.com/blog/baselines-acktr-a2c/)
        * [Proximal Policy Optimization (PPO)](https://arxiv.org/pdf/1707.06347.pdf)
        * [Soft Actor Critic (SAC)](https://arxiv.org/pdf/1812.05905.pdf)
        
        #### Multi-agent algorithms:
        * [Multi-agent DDPG (MADDPG)](https://arxiv.org/pdf/1706.02275.pdf)
        
        #### Massively parallel algorithms:
        * [Asynchronous A2C (A3C)](https://arxiv.org/abs/1602.01783)
        * [APEX-DQN, APEX-DDPG](https://arxiv.org/pdf/1803.00933)
        * [IMPALA](https://arxiv.org/pdf/1802.01561)
        
        #### Enhancements:
        * [Prioritized Experience Replay (PER)](https://arxiv.org/pdf/1511.05952.pdf)
        * [Generalized Advantage Estimation (GAE)](https://arxiv.org/pdf/1506.02438.pdf)
        * [Recurrent networks in DQN, etc.](https://arxiv.org/pdf/1507.06527.pdf)
        #### Algorithms to be supported:
        * [Generative Adversarial Imitation Learning (GAIL)](https://arxiv.org/abs/1606.03476)
        * Evolution Strategies
        * [QMIX (multi agent)](https://arxiv.org/abs/1803.11485)
        * Model-based methods
        
        ### Features
        ---
        #### 1. Readable
        
        Compared to other reinforcement learning libraries such as the famous [rlpyt](https://github.com/astooke/rlpyt), [ray](https://github.com/ray-project/ray), and [baselines](https://github.com/openai/baselines). Machin tries to just provide a simple, clear implementation of RL algorithms.
        
        All algorithms in Machin are designed with minimial abstractions and have very detailed documents, as well as various helpful tutorials.
        
        #### 2. Reusable
        
        Machin takes a similar approach to that of pytorch, encasulating algorithms, data structures in their own classes. Users do not need to setup a series of `data collectors`, `trainers`, `runners`, `samplers`... to use them, **just import**.
        
        The only restriction placed on your models is their input / output format, however, these restrictions are minimal, making it easy to adapt algorithms to your custom environments. 
        
        #### 3. Extendable
        Machin is built upon pytorch, it and thanks to its powerful rpc api, we may construct complex distributed programs. Machin provides implementations for enhanced parallel execution pools, automatic model assignment, role based rpc scaling, rpc service discovery and registration, etc.
        
        Upon these core functions, Machin is able to provide tested high-performance distributed training algorithm implementations, such as A3C, APEX, IMPALA, to ease your design.
        
        #### 4. Reproducible
        Machin is **weakly** reproducible, for each release, our test framework will directly train every RL framework, if any framework cannot reach the target score, the test will fail directly. 
        
        However, currently, the tests are not guaranteed to
        be exactly the same as the tests in original papers, due to the large variety of different environments used in original research papers.
        
        
        ### Documentation
        ---
        See [here](https://machin.readthedocs.io/). Examples are located in [examples](https://github.com/iffiX/machin/tree/master/examples).
        
        ### Installation
        ---
        Machin is hosted on [PyPI](https://pypi.org/project/machin/). Python >= 3.5 and PyTorch >= 1.5.0 is required. You may install the Machin library by simply typing:
        ```
        pip3 install machin
        ```
        
        ### Contributing
        ---
        Any contribution would be great, don't hesitate to submit a PR request to us! Please follow the instructions in [this](https://github.com/iffiX/machin/tree/master/docs/misc/contribute.md) file.
        
        ### Issues
        ---
        If you have any issues, please use the template markdown files in [.github/ISSUE_TEMPLATE](https://github.com/iffiX/machin/tree/master/.github/ISSUE_TEMPLATE) 
        folder and  format your issue before opening a new one. We would try our best to respond to your feature requests and problems.
        
        ### Citing
        ---
        We would be very grateful if you can cite our work in your publications:
        ```
        @misc{machin,
          author = {Muhan Li},
          title = {Machin},
          year = {2020},
          publisher = {GitHub},
          journal = {GitHub repository},
          howpublished = {\url{https://github.com/iffiX/machin}},
        }
        ```
        
        ### Roadmap
        ---
        Please see [Roadmap](https://github.com/iffiX/machin/projects/2) for the exciting new features we are currently working on!
        
Platform: UNKNOWN
Classifier: Development Status :: 4 - Beta
Classifier: Intended Audience :: Developers
Classifier: Intended Audience :: Science/Research
Classifier: Topic :: Scientific/Engineering :: Artificial Intelligence
Classifier: Topic :: Software Development :: Libraries :: Python Modules
Classifier: License :: OSI Approved :: MIT License
Classifier: Programming Language :: Python :: 3.5
Classifier: Programming Language :: Python :: 3.6
Classifier: Programming Language :: Python :: 3.7
Classifier: Programming Language :: Python :: 3.8
Requires-Python: >=3.5
Description-Content-Type: text/markdown
