ostatslib.agents.ppo_agent# PPO Agent module Classes PPOAgent([path, training_envs_count, ...]) Agent built on PPO algorithm model class PPOAgent(path: str | None = None, training_envs_count: int = 8, environment_kwargs: dict | None = None)[source]# Agent built on PPO algorithm model save(path: str) → None[source]# Saves agent prediction model Parameters: path (str) – path to file train(steps: int = 1000000, save_freq: int = 100000) → None[source]# Trains an agent Parameters: steps (int, optional) – Maximum number of steps during training. Defaults to 100e3.
