search

hugocen / freqtrade-gym

A customized gym environment for developing and comparing reinforcement learning algorithms in crypto trading.
GNU General Public License v3.0
223 stars 45 forks source link

The best backtests results #14

Closed xsa-dev closed 3 years ago

xsa-dev commented 3 years ago

Hello, buds

I am test this on Tensorflow version and my results of backtests is not good enough for dry run mode tests. You can see it attached file: ft_rl_results-default.txt

Whats your results?

hugocen commented 3 years ago

Yea, my results with deep reinforcement learning are not so good either. But I got some okish results using NEAT algorithms. We still need a lot of tunning and research to make this work.

evo11x commented 3 years ago

can you post the NEAT implementation with this strategy?

hugocen commented 3 years ago

@evo11x I updated the neat_trade.py for traing. And Neat_RL.py for loading the model. Sorry, but it's kinda sloppy.