Function watch_agent

rl/cartpole.py:85–94

(model, env, eps)

def watch_agent(model, env, eps):
  done = False
  truncated = False
  episode_reward = 0
  s, info = env.reset()
  while not (done or truncated):
    a = epsilon_greedy(model, s, eps=eps)
    s, r, done, truncated, info = env.step(a)
    episode_reward += r
  print("Episode reward:", episode_reward)


if __name__ == '__main__':