print(s)
env.close()
if close:
break
return score
if __name__ == "__main__":
# init ES
e = es.EvolutionStrategy(
outw,
1.0,
50, # much smaller than distributed
15,
min_sigma=1e-3,
big_sigma=5e-2,
wait_iter=100000)
# multiprocessing
pool = mp.Pool()
LENGTH = 1000
times = 0
best = -float('inf')
try:
for i in range(1000):
scores = []
pop = e.ask()
print(s)
env.close()
if close:
break
return score / n
if __name__ == "__main__":
# init ES
e = es.EvolutionStrategy(
outw,
1.0,
100,
10,
min_sigma=1e-3,
big_sigma=1e1,
wait_iter=5
)
# multiprocessing
pool = mp.Pool()
LENGTH = 1000
times = 0
best = -float('inf')
hist = open('car_es_hist.txt', 'w')
try:
print(s)
env.close()
if close:
break
return score / 3
if __name__ == "__main__":
# init ES
e = es.EvolutionStrategy(
outw,
1.0,
1000,
15,
min_sigma=1e-3,
big_sigma=5e-2,
wait_iter=100000
)
# distributed training
pool = distrib.DistributedServer()
pool.start()
print("Waiting for connections...")
time.sleep(5)
print("Done!")
LENGTH = 1000
times = 0