def cb(outputs):
try:
distrib1, distrib2, value1, value2 = outputs.result()
except CancelledError:
logger.info("Client {} cancelled.".format(ident))
return
assert np.all(np.isfinite(distrib1)), distrib1
assert np.all(np.isfinite(distrib2)), distrib2
rand_num = np.random.rand()
if rand_num < 0.5:
action = np.random.choice(len(distrib1), p=distrib1)
updateweight1, updateweight2 = 1.0, 0.0
else:
action = np.random.choice(len(distrib2), p=distrib2)
updateweight2, updateweight1 = 1.0, 0.0
client = self.clients[ident]
client.memory.append(
TransitionExperience(state,
action,
reward=None,
value1=value1,
value2=value2,
updateweight1=updateweight1,
updateweight2=updateweight2,
prob1=distrib1[action],
prob2=distrib2[action]))
self.send_queue.put([ident, dumps(action)])
def cb(outputs):
try:
distrib, value = outputs.result()
except CancelledError:
logger.info("Client {} cancelled.".format(client.ident))
return
assert np.all(np.isfinite(distrib)), distrib
action = np.random.choice(len(distrib), p=distrib)
client.memory.append(TransitionExperience(
state, action, reward=None, value=value, prob=distrib[action]))
self.send_queue.put([client.ident, dumps(action)])