def dora_run_softmax(run_name='default', plot=False):
selection = LLL_softmax()
env = CartPoleVision()
Qnet = selectNet()
Enet = selectNet(Enet=True)
t = DoraTrainer(Qnet, Enet, env, selection, lr=0.01, run_name=run_name, plot=plot)
t.run()
def optimistic_dqn_run_softmax(run_name='default', plot=False):
selection = softmax()
env = CartPoleVision()
optQnet = selectNet(Mode="optimisticQ")
t = Trainer(optQnet, env, selection, run_name=run_name, plot=plot)
t.run()
def dqn_run(run_name='default', plot=False):
selection = epsilon_greedy()
env = CartPoleVision()
Qnet = selectNet()
t = Trainer(Qnet, env, selection, run_name=run_name, plot=plot)
t.run()
def dqn_run_softmax(run_name='default', plot=False):
selection = softmax()
env = CartPoleVision()
Qnet = selectNet()
t = Trainer(Qnet, env, selection, run_name=run_name, plot=plot)
t.run()
def dora_run(env,
run_name='default',
plot=False,
selection='softmax',
setting=DefaultSetting(),
log_path='logs'):
selection = {
'epsilon': LLL_epsilon_greedy(),
'softmax': LLL_softmax()
}[selection]
Qnet = selectNet('dqn', env.name)
Enet = selectNet('enet', env.name)
t = DoraTrainer(Qnet,
Enet,
env,
selection,
run_name=run_name,
plot=plot,
setting=setting,
log_path=log_path)
t.run()