preprocess environment In OpenAI gym, it is class Wrapper(). In ElegantRL, it is a decorator What preprocess environment do? 1.1 use get_

Contribution of DQN (Deep Q Network) 1. Q-table (discrete state space) --> Q-network (continuous state space) 2. Use experiment replay

get information of a standard OpenAI gym env. The DRL algorithm AgentXXX need these env information for building networks and training. `obj

print line_intersection((A, B), (C, D)) line = ((x1, y1), (x2, y2)) How do I compute the intersection point of two lines? source: htt

main method to do the feature engineering @:param config: source dataframe @:return: a DataMatrices object

sample all the data in ReplayBuffer (for on-policy) :return torch.Tensor reward: reward.shape==(now_len, 1) :return torch.Tensor m

sample all the data in ReplayBuffer (for on-policy) :return torch.Tensor reward: reward.shape==(now_len, 1) :return torch.Tensor m

save or load model files `str cwd` current working directory, we save model file here `bool if_save` save model or load model