Agents

This folder includes 4 sorts of agent policies.

class mlp_agent.MLPAgent(input_shape, args) is the policy of MLPs with learned mean of a continuous action only.

class mlp_agent_gaussian.MLPAgent(input_shape, args) is the policy of MLPs with learned mean and log_std of a continuous action.

class rnn_agent.RNNAgent(input_shape, args) is the policy of GRUs with learned mean of a continuous action only.

class rnn_agent_gaussian.RNNAgent(input_shape, args) is the policy of GRUs with learned mean and log_std of a continuous action.