def __init__(self, mac, scheme, logger, args):
self.args = args
self.mac = mac
self.logger = logger
self.params = list(mac.parameters())
self.last_target_update_episode = 0
self.mixer = None
if args.mixer is not None:
if args.mixer == "vdn":
self.mixer = VDNMixer()
elif args.mixer == "qmix":
self.mixer = QMixer(args)
elif args.mixer == "qatten":
self.mixer = QattenMixer(args)
else:
raise ValueError("Mixer {} not recognised.".format(args.mixer))
self.params += list(self.mixer.parameters())
self.target_mixer = copy.deepcopy(self.mixer)
self.optimiser = RMSprop(params=self.params,
lr=args.lr,
alpha=args.optim_alpha,
eps=args.optim_eps)
# a little wasteful to deepcopy (e.g. duplicates action selector), but should work for any MAC
self.target_mac = copy.deepcopy(mac)
self.log_stats_t = -self.args.learner_log_interval - 1
def __init__(self, mac, scheme, logger, args):
self.args = args
self.mac = mac # 控制器
self.logger = logger
self.last_target_update_episode = 0
self.device = th.device('cuda' if args.use_cuda else 'cpu')
self.params = list(mac.parameters())
if args.mixer == "qatten":
self.mixer = QattenMixer(args)
elif args.mixer == "vdn":
self.mixer = VDNMixer()
elif args.mixer == "qmix":
self.mixer = Mixer(args)
else:
raise Exception("mixer error")
self.target_mixer = copy.deepcopy(self.mixer)
self.params += list(self.mixer.parameters())
print('Mixer的参数量为: ', end='')
print(get_parameters_num(self.mixer.parameters()))
if self.args.optimizer == 'adam':
self.optimiser = Adam(params=self.params, lr=args.lr)
else:
self.optimiser = RMSprop(params=self.params, lr=args.lr, alpha=args.optim_alpha, eps=args.optim_eps)
# 深度复制有点浪费（例如重复动作选择器），但对任何MAC都应该有效。
#
self.target_mac = copy.deepcopy(mac)
# 设置日志打印间隔
self.log_stats_t = -self.args.learner_log_interval - 1
self.train_t = 0
def __init__(self, mac, scheme, logger, args):
self.args = args
self.mac = mac
self.logger = logger
self.last_target_update_episode = 0
self.device = th.device('cuda' if args.use_cuda else 'cpu')
self.params = list(mac.parameters())
if args.mixer == "qatten":
self.mixer = QattenMixer(args)
elif args.mixer == "vdn":
self.mixer = VDNMixer()
elif args.mixer == "qmix":
self.mixer = Mixer(args)
else:
raise "mixer error"
self.target_mixer = copy.deepcopy(self.mixer)
self.params += list(self.mixer.parameters())
print('Mixer Size: ')
print(get_parameters_num(self.mixer.parameters()))
if self.args.optimizer == 'adam':
self.optimiser = Adam(params=self.params,
lr=args.lr,
weight_decay=getattr(args, "weight_decay",
0))
else:
self.optimiser = RMSprop(params=self.params,
lr=args.lr,
alpha=args.optim_alpha,
eps=args.optim_eps)
# a little wasteful to deepcopy (e.g. duplicates action selector), but should work for any MAC
self.target_mac = copy.deepcopy(mac)
self.log_stats_t = -self.args.learner_log_interval - 1
self.train_t = 0
# priority replay
self.use_per = getattr(self.args, 'use_per', False)
self.return_priority = getattr(self.args, "return_priority", False)
if self.use_per:
self.priority_max = float('-inf')
self.priority_min = float('inf')
class QattenLearner:
def __init__(self, mac, scheme, logger, args):
self.args = args
self.mac = mac
self.logger = logger
self.params = list(mac.parameters())
self.last_target_update_episode = 0
self.mixer = None
if args.mixer is not None:
if args.mixer == "vdn":
self.mixer = VDNMixer()
elif args.mixer == "qmix":
self.mixer = QMixer(args)
elif args.mixer == "qatten":
self.mixer = QattenMixer(args)
else:
raise ValueError("Mixer {} not recognised.".format(args.mixer))
self.params += list(self.mixer.parameters())
self.target_mixer = copy.deepcopy(self.mixer)
self.optimiser = RMSprop(params=self.params,
lr=args.lr,
alpha=args.optim_alpha,
eps=args.optim_eps)
# a little wasteful to deepcopy (e.g. duplicates action selector), but should work for any MAC
self.target_mac = copy.deepcopy(mac)
self.log_stats_t = -self.args.learner_log_interval - 1
def train(self,
batch: EpisodeBatch,
t_env: int,
episode_num: int,
show_demo=False,
save_data=None):
# Get the relevant quantities
rewards = batch["reward"][:, :-1]
actions = batch["actions"][:, :-1]
terminated = batch["terminated"][:, :-1].float()
mask = batch["filled"][:, :-1].float()
mask[:, 1:] = mask[:, 1:] * (1 - terminated[:, :-1])
avail_actions = batch["avail_actions"]
# Calculate estimated Q-Values
mac_out = []
self.mac.init_hidden(batch.batch_size)
for t in range(batch.max_seq_length):
agent_outs = self.mac.forward(batch, t=t)
mac_out.append(agent_outs)
mac_out = th.stack(mac_out, dim=1) # Concat over time
# Pick the Q-Values for the actions taken by each agent
chosen_action_qvals = th.gather(mac_out[:, :-1], dim=3,
index=actions).squeeze(
3) # Remove the last dim
x_mac_out = mac_out.clone().detach()
x_mac_out[avail_actions == 0] = -9999999
max_action_qvals, max_action_index = x_mac_out[:, :-1].max(dim=3)
max_action_index = max_action_index.detach().unsqueeze(3)
is_max_action = (max_action_index == actions).int().float()
if show_demo:
q_i_data = chosen_action_qvals.detach().cpu().numpy()
q_data = (max_action_qvals -
chosen_action_qvals).detach().cpu().numpy()
# Calculate the Q-Values necessary for the target
target_mac_out = []
self.target_mac.init_hidden(batch.batch_size)
for t in range(batch.max_seq_length):
target_agent_outs = self.target_mac.forward(batch, t=t)
target_mac_out.append(target_agent_outs)
# We don't need the first timesteps Q-Value estimate for calculating targets
target_mac_out = th.stack(target_mac_out[1:],
dim=1) # Concat across time
# Mask out unavailable actions
target_mac_out[avail_actions[:, 1:] == 0] = -9999999
# Max over target Q-Values
if self.args.double_q:
# Get actions that maximise live Q (for double q-learning)
mac_out_detach = mac_out.clone().detach()
mac_out_detach[avail_actions == 0] = -9999999
cur_max_actions = mac_out_detach[:, 1:].max(dim=3, keepdim=True)[1]
target_max_qvals = th.gather(target_mac_out, 3,
cur_max_actions).squeeze(3)
else:
target_max_qvals = target_mac_out.max(dim=3)[0]
cur_max_actions = target_mac_out.max(
dim=3, keepdim=True)[1] # get the indices
# cur_max_actions: (episode_batch, episode_length - 1, agent_num, 1)
target_next_actions = cur_max_actions.detach(
) # actions are also inputs for mixer network
# Mix
if self.mixer is not None:
if self.mixer.name == 'qatten':
chosen_action_qvals, q_attend_regs, head_entropies = self.mixer(
chosen_action_qvals, batch["state"][:, :-1], actions)
target_max_qvals, _, _ = self.target_mixer(
target_max_qvals, batch["state"][:, 1:],
target_next_actions)
else:
chosen_action_qvals = self.mixer(chosen_action_qvals,
batch["state"][:, :-1])
target_max_qvals = self.target_mixer(target_max_qvals,
batch["state"][:, 1:])
# Calculate 1-step Q-Learning targets
targets = rewards + self.args.gamma * (1 -
terminated) * target_max_qvals
if show_demo:
tot_q_data = chosen_action_qvals.detach().cpu().numpy()
tot_target = targets.detach().cpu().numpy()
print('action_pair_%d_%d' % (save_data[0], save_data[1]),
np.squeeze(q_data[:, 0]), np.squeeze(q_i_data[:, 0]),
np.squeeze(tot_q_data[:, 0]), np.squeeze(tot_target[:, 0]))
self.logger.log_stat(
'action_pair_%d_%d' % (save_data[0], save_data[1]),
np.squeeze(tot_q_data[:, 0]), t_env)
return
# Td-error
td_error = (chosen_action_qvals - targets.detach())
mask = mask.expand_as(td_error)
# 0-out the targets that came from padded data
masked_td_error = td_error * mask
# Normal L2 loss, take mean over actual data
if self.mixer.name == 'qatten':
loss = (masked_td_error**2).sum() / mask.sum() + q_attend_regs
else:
loss = (masked_td_error**2).sum() / mask.sum()
masked_hit_prob = th.mean(is_max_action, dim=2) * mask
hit_prob = masked_hit_prob.sum() / mask.sum()
# Optimise
self.optimiser.zero_grad()
loss.backward()
grad_norm = th.nn.utils.clip_grad_norm_(self.params,
self.args.grad_norm_clip)
self.optimiser.step()
if (episode_num - self.last_target_update_episode
) / self.args.target_update_interval >= 1.0:
self._update_targets()
self.last_target_update_episode = episode_num
if t_env - self.log_stats_t >= self.args.learner_log_interval:
self.logger.log_stat("loss", loss.item(), t_env)
self.logger.log_stat("hit_prob", hit_prob.item(), t_env)
self.logger.log_stat("grad_norm", grad_norm, t_env)
mask_elems = mask.sum().item()
self.logger.log_stat(
"td_error_abs",
(masked_td_error.abs().sum().item() / mask_elems), t_env)
self.logger.log_stat("q_taken_mean",
(chosen_action_qvals * mask).sum().item() /
(mask_elems * self.args.n_agents), t_env)
self.logger.log_stat("target_mean", (targets * mask).sum().item() /
(mask_elems * self.args.n_agents), t_env)
if self.mixer.name == 'qatten':
[
self.logger.log_stat('head_{}_entropy'.format(h_i),
ent.item(), t_env)
for h_i, ent in enumerate(head_entropies)
]
self.log_stats_t = t_env
def _update_targets(self):
self.target_mac.load_state(self.mac)
if self.mixer is not None:
self.target_mixer.load_state_dict(self.mixer.state_dict())
self.logger.console_logger.info("Updated target network")
def cuda(self):
self.mac.cuda()
self.target_mac.cuda()
if self.mixer is not None:
self.mixer.cuda()
self.target_mixer.cuda()
def save_models(self, path):
self.mac.save_models(path)
if self.mixer is not None:
th.save(self.mixer.state_dict(), "{}/mixer.th".format(path))
th.save(self.optimiser.state_dict(), "{}/opt.th".format(path))
def load_models(self, path):
self.mac.load_models(path)
# Not quite right but I don't want to save target networks
self.target_mac.load_models(path)
if self.mixer is not None:
self.mixer.load_state_dict(
th.load("{}/mixer.th".format(path),
map_location=lambda storage, loc: storage))
self.optimiser.load_state_dict(
th.load("{}/opt.th".format(path),
map_location=lambda storage, loc: storage))