Function get_agent

rl3/a2c/play.py:19–27

(env, nsteps=5, nstack=1, total_timesteps=int(80e6),
              vf_coef=0.5, ent_coef=0.01, max_grad_norm=0.5, lr=7e-4,
              epsilon=1e-5, alpha=0.99)

def get_agent(env, nsteps=5, nstack=1, total_timesteps=int(80e6),
              vf_coef=0.5, ent_coef=0.01, max_grad_norm=0.5, lr=7e-4,
              epsilon=1e-5, alpha=0.99):
    # Note: nstack=1 since frame_stack=True, during training frame_stack=False
    agent = Agent(Network=CNN, ob_space=env.observation_space,
                  ac_space=env.action_space, nenvs=1, nsteps=nsteps, nstack=nstack,
                  ent_coef=ent_coef, vf_coef=vf_coef, max_grad_norm=max_grad_norm,
                  lr=lr, alpha=alpha, epsilon=epsilon, total_timesteps=total_timesteps)
    return agent


def main():