search

ahq1993 / inverse_rl

Adversarial Imitation Via Variational Inverse Reinforcement Learning
MIT License
93 stars 11 forks source link

what means the max iteration of target empowerment model #3

Closed Usaywook closed 4 years ago

Usaywook commented 4 years ago

code at ant_irl.py

    empw_model = Empowerment(env=env,fusion=True, max_itrs=4)
    t_empw_model = Empowerment(env=env,scope='t_efn',fusion=True, max_itrs=2, name='empowerment2')

but we do not call t_empw_model.fit(...) why we set max_itrs=2 for target empowerment model? where is the target empowerment model updated?

ahq1993 commented 4 years ago

That must be a mistake, but you are right we do not use max_itrs for target model.