Metadata-Version: 2.1
Name: ai-traineree
Version: 0.0.14
Summary: Yet another zoo of (Deep) Reinforcment Learning methods in Python using PyTorch
Home-page: https://github.com/laszukdawid/ai-traineree
Author: Dawid Laszuk
Author-email: ai-traineree@dawid.lasz.uk
License: UNKNOWN
Platform: UNKNOWN
Classifier: Programming Language :: Python :: 3
Classifier: License :: OSI Approved :: Apache Software License
Classifier: Operating System :: OS Independent
Classifier: Intended Audience :: Developers
Classifier: Intended Audience :: Science/Research
Classifier: Topic :: Scientific/Engineering :: Artificial Intelligence
Requires-Python: >=3.6
Description-Content-Type: text/markdown
Provides-Extra: plot
Provides-Extra: doc
Provides-Extra: test
Provides-Extra: tensorboard
Provides-Extra: neptune
Provides-Extra: gym
Provides-Extra: mlagents
Provides-Extra: pettingzoo

# ai-traineree
[![DocStatus](https://readthedocs.org/projects/ai-traineree/badge/?version=latest)](https://ai-traineree.readthedocs.io/)
[![Build Status](https://travis-ci.com/laszukdawid/ai-traineree.svg?branch=master)](https://travis-ci.com/laszukdawid/ai-traineree)
[![codecov](https://codecov.io/gh/laszukdawid/ai-traineree/branch/master/graph/badge.svg?token=S62DK7HPYA)](https://codecov.io/gh/laszukdawid/ai-traineree)

The intention is to have a zoo of Deep Reinforcment Learning methods and showcasing their application on some environments.

Read more in the doc: [ReadTheDocs AI-Traineree](https://ai-traineree.readthedocs.io/).

## Why another?

The main reason is the implemention philosophy.
We strongly believe that agents should be emerged in the environment and not the other way round.
Majority of the popular implementations pass environment instance to the agent as if the agent was the focus point.
This might ease implementation of some algorithms but it isn't representative of the world;
agents want to control the environment but that doesn't mean they can/should.

That, and using PyTorch instead of Tensorflow or JAX.

## Quick start

To get started with training your RL agent you need three things: an agent, an environment and a runner. Let's say you want to train a DQN agent on OpenAI CartPole-v1:
```
from ai_traineree.agents.dqn import DQNAgent
from ai_traineree.env_runner import EnvRunner
from ai_traineree.tasks import GymTask

task = GymTask('CartPole-v1)
agent = DQNAgent(task.state_size, task.action_size)
env_runner = EnvRunner(task, agent)

scores = env_runner.run()
```
or execute one of provided examples
```
>  python -m examples.cart_dqn
```

That's it.

## Installation

There isn't currently any installation mechanism. Git clone is expected if you want to play yourself. Coming updates include pip package and installation instructions.

As usual with Python, the expectation is to have own virtual environment and then pip install requirements. For example,
```bash
> python -m venv .venv
> git clone git@github.com:laszukdawid/ai-traineree.git
> source .venv/bin/activate
> python setup.py install
```

## Current state

### Playing gym
One way to improve learning speed is to simply show them how to play or, more researchy/creepy, provide a proper seed.
This isn't a general rule, since some algorithms train better without any human interaction, but since you're on GitHub... that's unlikely your case.
Currently there's a script [`interact.py`](scripts/interact.py) which uses OpenAI Gym's play API to record moves and AI Traineree to store them
in a buffer. Such buffers can be loaded by agents on initiation.

This is just a beginning and there will be more work on these interactions.

*Requirement*: Install `pygame`.

### Agents

| Short | Progress | Link | Full name | Doc |
|-------|----------|------|-----------|-----|
| DQN  | [Implemented](ai_traineree/agents/dqn.py) | [DeepMind](https://deepmind.com/research/publications/human-level-control-through-deep-reinforcement-learning), [Nature](https://www.nature.com/articles/nature14236)| Deep Q-learning Network  | [Doc](https://ai-traineree.readthedocs.io/en/latest/agents.html#dqn)
| DDPG | [Implemented](ai_traineree/agents/ddpg.py) | [arXiv](https://arxiv.org/abs/1509.02971) | Deep Deterministic Policy Gradient | [Doc](https://ai-traineree.readthedocs.io/en/latest/agents.html#ddpg)
| D4PG | [Implemented](ai_traineree/agents/d4pg.py) | [arXiv](https://arxiv.org/abs/1804.08617) | Distributed Distributional Deterministic Policy Gradients | [Doc](https://ai-traineree.readthedocs.io/en/latest/agents.html#d4pg)
| TD3 | [Implemented](ai_traineree/agents/td3.py) | [arXiv](https://arxiv.org/abs/1802.09477) | Twine Delayed Deep Deterministic policy gradient | [Doc](https://ai-traineree.readthedocs.io/en/latest/agents.html#td3)
| PPO | [Implemented](ai_traineree/agents/ppo.py) | [arXiv](https://arxiv.org/abs/1707.06347) | Proximal Policy Optimization |[Doc](https://ai-traineree.readthedocs.io/en/latest/agents.html#ppo)
| SAC | [Implemented](ai_traineree/agents/sac.py) | [arXiv](https://arxiv.org/abs/1801.01290) | Soft Actor Critic |[Doc](https://ai-traineree.readthedocs.io/en/latest/agents.html#sac)
| TRPO | | [arXiv](https://arxiv.org/abs/1502.05477) | Trust Region Policy Optimization |
| RAINBOW | [Implemented](ai_traineree/agents/rainbow.py) | [arXiv](https://arxiv.org/abs/1710.02298) | DQN with a few improvements |[Doc](https://ai-traineree.readthedocs.io/en/latest/agents.html#rainbow)

### Multi agents

We provide both Multi Agents agents entities and means to execute them against supported (below) environements.
However, that doesn't mean one can be used without the other.

| Short | Progress | Link | Full name | Doc |
|-------|----------|------|-----------|-----|
| IQL | [Implemented](ai_traineree/multi_agent/iql.py) | | Independent Q-Learners|[Doc](https://ai-traineree.readthedocs.io/en/latest/multi_agent.html#iql)
| MADDPG | [Implemented](ai_traineree/multi_agent/maddpg.py) | [arXiv](https://arxiv.org/abs/1706.02275) | Multi agent DDPG |[Doc](https://ai-traineree.readthedocs.io/en/latest/multi_agent.html#maddpg)

### Loggers

Supports using Tensorboard (via PyTorch's [SummaryWriter](https://pytorch.org/docs/stable/tensorboard.html)) and [Neptune](https://neptune.ai) to display metrics. Wrappers are provided as `TensorboardLogger` and `NeptuneLogger`.

*Note*: In order to use Neptune one needs to install [neptune-client](https://pypi.org/project/neptune-client/) (`pip install neptune-client`).

### Environments

| Name | Progress | Link |
|------|----------|------|
| OpenAI Gym - Classic | Done |
| OpenAI Gym - Atari | Done |
| OpenAI Gym - MuJoCo | Not interested. |
| PettingZoo | Initial support | [Page](https://www.pettingzoo.ml/) / [GitHub](https://github.com/PettingZoo-Team/PettingZoo) |
| Unity ML | Somehow supported. | [Page](https://unity3d.com/machine-learning)
| MAME Linux emulator | Interested. | [Official page](https://www.mamedev.org/)

### Development

| Name | Progress |
|------|----------|
| CD | Not started |
| More multi agent methods | Research |
| Test coverage > 80% | Tested ~40%, Covered 85% |

There are other potential things on the roadmap but haven't dedicated to them yet. 

Should I focus on something specificallly? Let me know by leaving opening a feature request issue or contacting through [ai-traineree@dawid.lasz.uk](mailto:ai-traineree@dawid.lasz.uk).



