notadamking / RLTrader

A cryptocurrency trading environment using deep reinforcement learning and OpenAI's gym
https://discord.gg/ZZ7BGWh
GNU General Public License v3.0
1.74k stars 538 forks source link

RLTrader — The Predecessor to TensorTrade

Build Status Contributions Welcome GPL License Discord Python 3.6 Github Release

Development on this library has slowed down, in favor of working on TensorTrade - a framework for trading with RL: https://github.com/notadamking/tensortrade

If you'd like to learn more about how we created this agent, check out the Medium article: https://towardsdatascience.com/creating-bitcoin-trading-bots-that-dont-lose-money-2e7165fb0b29

Later, we optimized this repo using feature engineering, statistical modeling, and Bayesian optimization, check it out: https://towardsdatascience.com/using-reinforcement-learning-to-trade-bitcoin-for-massive-profit-b69d0e8f583b

Discord server: https://discord.gg/ZZ7BGWh

Data sets: https://www.cryptodatadownload.com/data/northamerican/

Live trading visualization

Getting Started

How to find out if you have nVIDIA GPU?

Linux:

sudo lspci | grep -i --color 'vga\|3d\|2d' | grep -i nvidia

If this returns anything, then you should have an nVIDIA card.

Basic usage

The first thing you will need to do to get started is install the requirements. If your system has an nVIDIA GPU that you should start by using:

cd "path-of-your-cloned-rl-trader-dir"
pip install -r requirements.txt

More information regarding how you can take advantage of your GPU while using docker: https://github.com/NVIDIA/nvidia-docker

If you have another type of GPU or you simply want to use your CPU, use:

pip install -r requirements.no-gpu.txt

Update your current static files, that are used by default:

 python ./cli.py update-static-data

Afterwards you can simply see the currently available options:

python ./cli.py --help

or simply run the project with default options:

python ./cli.py optimize

If you have a standard set of configs you want to run the trader against, you can specify a config file to load configuration from. Rename config/config.ini.dist to config/config.ini and run

python ./cli.py --from-config config/config.ini optimize
python ./cli.py optimize

Testing with vagrant

Start the vagrant box using:

vagrant up

Code will be located at /vagrant. Play and/or test with whatever package you wish. Note: With vagrant you cannot take full advantage of your GPU, so is mainly for testing purposes

Testing with docker

If you want to run everything within a docker container, then just use:

./run-with-docker (cpu|gpu) (yes|no) optimize
python ./ cli.py --params-db-path "postgres://rl_trader:rl_trader@localhost" optimize

The database and it's data are pesisted under data/postgres locally.

If you want to spin a docker test environment:

./run-with-docker (cpu|gpu) (yes|no)

If you want to run existing tests, then just use:

./run-tests-with-docker

Fire up a local docker dev environment

./dev-with-docker

Windows 10 installation, no CUDA installation needed

conda create --name rltrader python=3.6.8 pip git conda activate rltrader conda install tensorflow-gpu git clone https://github.com/notadamking/RLTrader pip install -r RLTrader/requirements.txt

Optimizing, Training, and Testing

While you could just let the agent train and run with the default PPO2 hyper-parameters, your agent would likely not be very profitable. The stable-baselines library provides a great set of default parameters that work for most problem domains, but we need to better.

To do this, you will need to run optimize.py.

python ./optimize.py

This can take a while (hours to days depending on your hardware setup), but over time it will print to the console as trials are completed. Once a trial is completed, it will be stored in ./data/params.db, an SQLite database, from which we can pull hyper-parameters to train our agent.

From there, agents will be trained using the best set of hyper-parameters, and later tested on completely new data to verify the generalization of the algorithm.

Feel free to ask any questions in the Discord!

Google Colab

Enter and run the following snippet in the first cell to load RLTrader into a Google Colab environment. Don't forget to set hardware acceleration to GPU to speed up training!

!git init && git remote add origin https://github.com/notadamking/RLTrader.git && git pull origin master
!pip install -r requirements.txt

Common troubleshooting

The specified module could not be found.

Normally this is caused by missing mpi module. You should install it according to your platorm.

Contributing

Contributions are encouraged and I will always do my best to get them implemented into the library ASAP. This project is meant to grow as the community around it grows. Let me know if there is anything that you would like to see in the future or if there is anything you feel is missing.

Working on your first Pull Request? You can learn how from this free series How to Contribute to an Open Source Project on GitHub

https://github.com/notadamking/rltrader/graphs/contributors

Support

Want to show your support for this project and help it grow?

Head over to the successor framework: https://github.com/notadamking/tensortrade

Supporters: