class SAC(object):
def __init__(self, num_inputs, action_space, args):
self.num_inputs = num_inputs
self.action_space = action_space.shape[0]
self.gamma = args.gamma
self.tau = args.tau
self.policy_type = args.policy
self.target_update_interval = args.target_update_interval
self.automatic_entropy_tuning = args.automatic_entropy_tuning
self.critic = QNetwork(self.num_inputs, self.action_space,
args.hidden_size)
self.critic_optim = Adam(self.critic.parameters(), lr=args.lr)
if self.policy_type == "Gaussian":
self.alpha = args.alpha
# Target Entropy = −dim(A) (e.g. , -6 for HalfCheetah-v2) as given in the paper
if self.automatic_entropy_tuning == True:
self.target_entropy = -torch.prod(
torch.Tensor(action_space.shape)).item()
self.log_alpha = torch.zeros(1, requires_grad=True)
self.alpha_optim = Adam([self.log_alpha], lr=args.lr)
else:
pass
self.policy = GaussianPolicy(self.num_inputs, self.action_space,
args.hidden_size)
self.policy_optim = Adam(self.policy.parameters(), lr=args.lr)
self.value = ValueNetwork(self.num_inputs, args.hidden_size)
self.value_target = ValueNetwork(self.num_inputs, args.hidden_size)
self.value_optim = Adam(self.value.parameters(), lr=args.lr)
hard_update(self.value_target, self.value)
else:
self.policy = DeterministicPolicy(self.num_inputs,
self.action_space,
args.hidden_size)
self.policy_optim = Adam(self.policy.parameters(), lr=args.lr)
self.critic_target = QNetwork(self.num_inputs, self.action_space,
args.hidden_size)
hard_update(self.critic_target, self.critic)
def select_action(self, state, eval=False):
state = torch.FloatTensor(state).unsqueeze(0)
if eval == False:
self.policy.train()
action, _, _, _, _ = self.policy.sample(state)
else:
self.policy.eval()
_, _, _, action, _ = self.policy.sample(state)
if self.policy_type == "Gaussian":
action = torch.tanh(action)
else:
pass
#action = torch.tanh(action)
action = action.detach().cpu().numpy()
return action[0]
def update_parameters(self, state_batch, action_batch, reward_batch,
next_state_batch, mask_batch, updates):
state_batch = torch.FloatTensor(state_batch)
next_state_batch = torch.FloatTensor(next_state_batch)
action_batch = torch.FloatTensor(action_batch)
reward_batch = torch.FloatTensor(reward_batch).unsqueeze(1)
mask_batch = torch.FloatTensor(np.float32(mask_batch)).unsqueeze(1)
"""
Use two Q-functions to mitigate positive bias in the policy improvement step that is known
to degrade performance of value based methods. Two Q-functions also significantly speed
up training, especially on harder task.
"""
expected_q1_value, expected_q2_value = self.critic(
state_batch, action_batch)
new_action, log_prob, _, mean, log_std = self.policy.sample(
state_batch)
if self.policy_type == "Gaussian":
if self.automatic_entropy_tuning:
"""
Alpha Loss
"""
alpha_loss = -(
self.log_alpha *
(log_prob + self.target_entropy).detach()).mean()
self.alpha_optim.zero_grad()
alpha_loss.backward()
self.alpha_optim.step()
self.alpha = self.log_alpha.exp()
alpha_logs = self.alpha.clone() # For TensorboardX logs
else:
alpha_loss = torch.tensor(0.)
alpha_logs = self.alpha # For TensorboardX logs
"""
Including a separate function approximator for the soft value can stabilize training.
"""
expected_value = self.value(state_batch)
target_value = self.value_target(next_state_batch)
next_q_value = reward_batch + mask_batch * self.gamma * (
target_value).detach()
else:
"""
There is no need in principle to include a separate function approximator for the state value.
We use a target critic network for deterministic policy and eradicate the value value network completely.
"""
alpha_loss = torch.tensor(0.)
alpha_logs = self.alpha # For TensorboardX logs
next_state_action, _, _, _, _, = self.policy.sample(
next_state_batch)
target_critic_1, target_critic_2 = self.critic_target(
next_state_batch, next_state_action)
target_critic = torch.min(target_critic_1, target_critic_2)
next_q_value = reward_batch + mask_batch * self.gamma * (
target_critic).detach()
"""
Soft Q-function parameters can be trained to minimize the soft Bellman residual
JQ = 𝔼(st,at)~D[0.5(Q1(st,at) - r(st,at) - γ(𝔼st+1~p[V(st+1)]))^2]
∇JQ = ∇Q(st,at)(Q(st,at) - r(st,at) - γV(target)(st+1))
"""
q1_value_loss = F.mse_loss(expected_q1_value, next_q_value)
q2_value_loss = F.mse_loss(expected_q2_value, next_q_value)
q1_new, q2_new = self.critic(state_batch, new_action)
expected_new_q_value = torch.min(q1_new, q2_new)
if self.policy_type == "Gaussian":
"""
Including a separate function approximator for the soft value can stabilize training and is convenient to
train simultaneously with the other networks
Update the V towards the min of two Q-functions in order to reduce overestimation bias from function approximation error.
JV = 𝔼st~D[0.5(V(st) - (𝔼at~π[Qmin(st,at) - α * log π(at|st)]))^2]
∇JV = ∇V(st)(V(st) - Q(st,at) + (α * logπ(at|st)))
"""
next_value = expected_new_q_value - (self.alpha * log_prob)
value_loss = F.mse_loss(expected_value, next_value.detach())
else:
pass
"""
Reparameterization trick is used to get a low variance estimator
f(εt;st) = action sampled from the policy
εt is an input noise vector, sampled from some fixed distribution
Jπ = 𝔼st∼D,εt∼N[α * logπ(f(εt;st)|st) − Q(st,f(εt;st))]
∇Jπ = ∇log π + ([∇at (α * logπ(at|st)) − ∇at Q(st,at)])∇f(εt;st)
"""
policy_loss = ((self.alpha * log_prob) - expected_new_q_value).mean()
# Regularization Loss
mean_loss = 0.001 * mean.pow(2).mean()
std_loss = 0.001 * log_std.pow(2).mean()
policy_loss += mean_loss + std_loss
self.critic_optim.zero_grad()
q1_value_loss.backward()
self.critic_optim.step()
self.critic_optim.zero_grad()
q2_value_loss.backward()
self.critic_optim.step()
if self.policy_type == "Gaussian":
self.value_optim.zero_grad()
value_loss.backward()
self.value_optim.step()
else:
value_loss = torch.tensor(0.)
self.policy_optim.zero_grad()
policy_loss.backward()
self.policy_optim.step()
"""
We update the target weights to match the current value function weights periodically
Update target parameter after every n(args.target_update_interval) updates
"""
if updates % self.target_update_interval == 0 and self.policy_type == "Deterministic":
soft_update(self.critic_target, self.critic, self.tau)
elif updates % self.target_update_interval == 0 and self.policy_type == "Gaussian":
soft_update(self.value_target, self.value, self.tau)
return value_loss.item(), q1_value_loss.item(), q2_value_loss.item(
), policy_loss.item(), alpha_loss.item(), alpha_logs
# Save model parameters
def save_model(self,
env_name,
suffix="",
actor_path=None,
critic_path=None,
value_path=None):
if not os.path.exists('models/'):
os.makedirs('models/')
if actor_path is None:
actor_path = "models/sac_actor_{}_{}".format(env_name, suffix)
if critic_path is None:
critic_path = "models/sac_critic_{}_{}".format(env_name, suffix)
if value_path is None:
value_path = "models/sac_value_{}_{}".format(env_name, suffix)
print('Saving models to {}, {} and {}'.format(actor_path, critic_path,
value_path))
torch.save(self.value.state_dict(), value_path)
torch.save(self.policy.state_dict(), actor_path)
torch.save(self.critic.state_dict(), critic_path)
# Load model parameters
def load_model(self, actor_path, critic_path, value_path):
print('Loading models from {}, {} and {}'.format(
actor_path, critic_path, value_path))
if actor_path is not None:
self.policy.load_state_dict(torch.load(actor_path))
if critic_path is not None:
self.critic.load_state_dict(torch.load(critic_path))
if value_path is not None:
self.value.load_state_dict(torch.load(value_path))
class SAC(object):
def __init__(self, num_inputs, action_space, args):
self.num_inputs = num_inputs
self.action_space = action_space.shape[0]
self.gamma = args.gamma
self.tau = args.tau
self.policy_type = args.policy
self.target_update_interval = args.target_update_interval
self.automatic_entropy_tuning = args.automatic_entropy_tuning
self.device = torch.device("cuda" if args.cuda else "cpu")
self.critic = QNetwork(self.num_inputs, self.action_space,
args.hidden_size).to(device=self.device)
self.critic_optim = Adam(self.critic.parameters(), lr=args.lr)
if self.policy_type == "Gaussian":
self.alpha = args.alpha
# Target Entropy = −dim(A) (e.g. , -6 for HalfCheetah-v2) as given in the paper
if self.automatic_entropy_tuning == True:
self.target_entropy = -torch.prod(
torch.Tensor(action_space.shape).to(self.device)).item()
self.log_alpha = torch.zeros(1,
requires_grad=True,
device=self.device)
self.alpha_optim = Adam([self.log_alpha], lr=args.lr)
self.policy = GaussianPolicy(self.num_inputs, self.action_space,
args.hidden_size).to(self.device)
self.policy_optim = Adam(self.policy.parameters(), lr=args.lr)
self.value = ValueNetwork(self.num_inputs,
args.hidden_size).to(self.device)
self.value_target = ValueNetwork(self.num_inputs,
args.hidden_size).to(self.device)
self.value_optim = Adam(self.value.parameters(), lr=args.lr)
hard_update(self.value_target, self.value)
else:
self.policy = DeterministicPolicy(self.num_inputs,
self.action_space,
args.hidden_size).to(self.device)
self.policy_optim = Adam(self.policy.parameters(), lr=args.lr)
self.critic_target = QNetwork(self.num_inputs, self.action_space,
args.hidden_size).to(self.device)
hard_update(self.critic_target, self.critic)
def select_action(self, state, eval=False):
state = torch.FloatTensor(state).to(self.device).unsqueeze(0)
if eval == False:
self.policy.train()
action, _, _ = self.policy.sample(state)
else:
self.policy.eval()
_, _, action = self.policy.sample(state)
action = action.detach().cpu().numpy()
return action[0]
def update_parameters(self, state_batch, action_batch, reward_batch,
next_state_batch, mask_batch, updates):
state_batch = torch.FloatTensor(state_batch).to(self.device)
next_state_batch = torch.FloatTensor(next_state_batch).to(self.device)
action_batch = torch.FloatTensor(action_batch).to(self.device)
reward_batch = torch.FloatTensor(reward_batch).to(
self.device).unsqueeze(1)
mask_batch = torch.FloatTensor(mask_batch).to(self.device).unsqueeze(1)
qf1, qf2 = self.critic(
state_batch, action_batch
) # Two Q-functions to mitigate positive bias in the policy improvement step
pi, log_pi, _ = self.policy.sample(state_batch)
if self.policy_type == "Gaussian":
if self.automatic_entropy_tuning:
alpha_loss = -(self.log_alpha *
(log_pi + self.target_entropy).detach()).mean()
self.alpha_optim.zero_grad()
alpha_loss.backward()
self.alpha_optim.step()
self.alpha = self.log_alpha.exp()
alpha_logs = torch.tensor(self.alpha) # For TensorboardX logs
else:
alpha_loss = torch.tensor(0.).to(self.device)
alpha_logs = torch.tensor(self.alpha) # For TensorboardX logs
vf = self.value(
state_batch
) # separate function approximator for the soft value can stabilize training.
with torch.no_grad():
vf_next_target = self.value_target(next_state_batch)
next_q_value = reward_batch + mask_batch * self.gamma * (
vf_next_target)
else:
alpha_loss = torch.tensor(0.).to(self.device)
alpha_logs = self.alpha # For TensorboardX logs
with torch.no_grad():
next_state_action, _, _, _, _, = self.policy.sample(
next_state_batch)
# Use a target critic network for deterministic policy and eradicate the value value network completely.
qf1_next_target, qf2_next_target = self.critic_target(
next_state_batch, next_state_action)
min_qf_next_target = torch.min(qf1_next_target,
qf2_next_target)
next_q_value = reward_batch + mask_batch * self.gamma * (
min_qf_next_target)
qf1_loss = F.mse_loss(
qf1, next_q_value
) # JQ = 𝔼(st,at)~D[0.5(Q1(st,at) - r(st,at) - γ(𝔼st+1~p[V(st+1)]))^2]
qf2_loss = F.mse_loss(
qf2, next_q_value
) # JQ = 𝔼(st,at)~D[0.5(Q1(st,at) - r(st,at) - γ(𝔼st+1~p[V(st+1)]))^2]
qf1_pi, qf2_pi = self.critic(state_batch, pi)
min_qf_pi = torch.min(qf1_pi, qf2_pi)
if self.policy_type == "Gaussian":
vf_target = min_qf_pi - (self.alpha * log_pi)
value_loss = F.mse_loss(
vf, vf_target.detach()
) # JV = 𝔼st~D[0.5(V(st) - (𝔼at~π[Qmin(st,at) - α * log π(at|st)]))^2]
policy_loss = ((self.alpha * log_pi) - min_qf_pi).mean(
) # Jπ = 𝔼st∼D,εt∼N[α * logπ(f(εt;st)|st) − Q(st,f(εt;st))]
# Regularization Loss
# mean_loss = 0.001 * mean.pow(2).mean()
# std_loss = 0.001 * log_std.pow(2).mean()
# policy_loss += mean_loss + std_loss
self.critic_optim.zero_grad()
qf1_loss.backward()
self.critic_optim.step()
self.critic_optim.zero_grad()
qf2_loss.backward()
self.critic_optim.step()
if self.policy_type == "Gaussian":
self.value_optim.zero_grad()
value_loss.backward()
self.value_optim.step()
else:
value_loss = torch.tensor(0.).to(self.device)
self.policy_optim.zero_grad()
policy_loss.backward()
self.policy_optim.step()
"""
We update the target weights to match the current value function weights periodically
Update target parameter after every n(args.target_update_interval) updates
"""
if updates % self.target_update_interval == 0 and self.policy_type == "Deterministic":
soft_update(self.critic_target, self.critic, self.tau)
elif updates % self.target_update_interval == 0 and self.policy_type == "Gaussian":
soft_update(self.value_target, self.value, self.tau)
return value_loss.item(), qf1_loss.item(), qf2_loss.item(
), policy_loss.item(), alpha_loss.item(), alpha_logs.item()
# Save model parameters
def save_model(self,
env_name,
suffix="",
actor_path=None,
critic_path=None,
value_path=None):
if not os.path.exists('models/'):
os.makedirs('models/')
if actor_path is None:
actor_path = "models/sac_actor_{}_{}".format(env_name, suffix)
if critic_path is None:
critic_path = "models/sac_critic_{}_{}".format(env_name, suffix)
if value_path is None:
value_path = "models/sac_value_{}_{}".format(env_name, suffix)
print('Saving models to {}, {} and {}'.format(actor_path, critic_path,
value_path))
torch.save(self.value.state_dict(), value_path)
torch.save(self.policy.state_dict(), actor_path)
torch.save(self.critic.state_dict(), critic_path)
# Load model parameters
def load_model(self, actor_path, critic_path, value_path):
print('Loading models from {}, {} and {}'.format(
actor_path, critic_path, value_path))
if actor_path is not None:
self.policy.load_state_dict(torch.load(actor_path))
if critic_path is not None:
self.critic.load_state_dict(torch.load(critic_path))
if value_path is not None:
self.value.load_state_dict(torch.load(value_path))
class SAC(object):
def __init__(self, num_inputs, action_space, args):
self.gamma = args.gamma
self.tau = args.tau
self.alpha = args.alpha
self.policy_type = args.policy
self.target_update_interval = args.target_update_interval
self.automatic_entropy_tuning = args.automatic_entropy_tuning
self.device = torch.device("cuda" if args.cuda else "cpu")
self.critic = QNetwork(num_inputs, action_space.shape[0],
args.hidden_size).to(device=self.device)
self.critic_optim = Adam(self.critic.parameters(), lr=args.lr)
self.critic_target = QNetwork(num_inputs, action_space.shape[0],
args.hidden_size).to(self.device)
hard_update(self.critic_target, self.critic)
if self.policy_type == "Gaussian":
# Target Entropy = −dim(A) (e.g. , -6 for HalfCheetah-v2) as given in the paper
if self.automatic_entropy_tuning is True:
self.target_entropy = -torch.prod(
torch.Tensor(action_space.shape).to(self.device)).item()
self.log_alpha = torch.zeros(1,
requires_grad=True,
device=self.device)
self.alpha_optim = Adam([self.log_alpha], lr=args.lr)
self.policy = GaussianPolicy(num_inputs, action_space.shape[0],
args.hidden_size,
action_space).to(self.device)
self.policy_optim = Adam(self.policy.parameters(), lr=args.lr)
else:
self.alpha = 0
self.automatic_entropy_tuning = False
self.policy = DeterministicPolicy(num_inputs,
action_space.shape[0],
args.hidden_size,
action_space).to(self.device)
self.policy_optim = Adam(self.policy.parameters(), lr=args.lr)
def select_action(self, state, evaluate=False):
state = torch.FloatTensor(state).to(self.device).unsqueeze(0)
if evaluate is False:
action, _, _ = self.policy.sample(state)
else:
_, _, action = self.policy.sample(state)
return action.detach().cpu().numpy()[0]
def update_parameters(self, memory, batch_size, updates):
# Sample a batch from memory
state_batch, action_batch, reward_batch, next_state_batch, mask_batch = memory.sample(
batch_size=batch_size)
state_batch = torch.FloatTensor(state_batch).to(self.device)
next_state_batch = torch.FloatTensor(next_state_batch).to(self.device)
action_batch = torch.FloatTensor(action_batch).to(self.device)
reward_batch = torch.FloatTensor(reward_batch).to(
self.device).unsqueeze(1)
mask_batch = torch.FloatTensor(mask_batch).to(self.device).unsqueeze(1)
with torch.no_grad():
next_state_action, next_state_log_pi, _ = self.policy.sample(
next_state_batch)
qf1_next_target, qf2_next_target = self.critic_target(
next_state_batch, next_state_action)
min_qf_next_target = torch.min(
qf1_next_target,
qf2_next_target) - self.alpha * next_state_log_pi
next_q_value = reward_batch + mask_batch * self.gamma * (
min_qf_next_target)
qf1, qf2 = self.critic(
state_batch, action_batch
) # Two Q-functions to mitigate positive bias in the policy improvement step
qf1_loss = F.mse_loss(
qf1, next_q_value
) # JQ = 𝔼(st,at)~D[0.5(Q1(st,at) - r(st,at) - γ(𝔼st+1~p[V(st+1)]))^2]
qf2_loss = F.mse_loss(
qf2, next_q_value
) # JQ = 𝔼(st,at)~D[0.5(Q1(st,at) - r(st,at) - γ(𝔼st+1~p[V(st+1)]))^2]
qf_loss = qf1_loss + qf2_loss
self.critic_optim.zero_grad()
qf_loss.backward()
self.critic_optim.step()
pi, log_pi, _ = self.policy.sample(state_batch)
qf1_pi, qf2_pi = self.critic(state_batch, pi)
min_qf_pi = torch.min(qf1_pi, qf2_pi)
policy_loss = ((self.alpha * log_pi) - min_qf_pi).mean(
) # Jπ = 𝔼st∼D,εt∼N[α * logπ(f(εt;st)|st) − Q(st,f(εt;st))]
self.policy_optim.zero_grad()
policy_loss.backward()
self.policy_optim.step()
if self.automatic_entropy_tuning:
alpha_loss = -(self.log_alpha *
(log_pi + self.target_entropy).detach()).mean()
self.alpha_optim.zero_grad()
alpha_loss.backward()
self.alpha_optim.step()
self.alpha = self.log_alpha.exp()
alpha_tlogs = self.alpha.clone() # For TensorboardX logs
else:
alpha_loss = torch.tensor(0.).to(self.device)
alpha_tlogs = torch.tensor(self.alpha) # For TensorboardX logs
if updates % self.target_update_interval == 0:
soft_update(self.critic_target, self.critic, self.tau)
return qf1_loss.item(), qf2_loss.item(), policy_loss.item(
), alpha_loss.item(), alpha_tlogs.item()
# Save model parameters
def save_checkpoint(self, env_name, suffix="", ckpt_path=None):
if not os.path.exists('checkpoints/'):
os.makedirs('checkpoints/')
if ckpt_path is None:
ckpt_path = "checkpoints/sac_checkpoint_{}_{}".format(
env_name, suffix)
print('Saving models to {}'.format(ckpt_path))
torch.save(
{
'policy_state_dict': self.policy.state_dict(),
'critic_state_dict': self.critic.state_dict(),
'critic_target_state_dict': self.critic_target.state_dict(),
'critic_optimizer_state_dict': self.critic_optim.state_dict(),
'policy_optimizer_state_dict': self.policy_optim.state_dict()
}, ckpt_path)
# Load model parameters
def load_checkpoint(self, ckpt_path, evaluate=False):
print('Loading models from {}'.format(ckpt_path))
if ckpt_path is not None:
checkpoint = torch.load(ckpt_path)
self.policy.load_state_dict(checkpoint['policy_state_dict'])
self.critic.load_state_dict(checkpoint['critic_state_dict'])
self.critic_target.load_state_dict(
checkpoint['critic_target_state_dict'])
self.critic_optim.load_state_dict(
checkpoint['critic_optimizer_state_dict'])
self.policy_optim.load_state_dict(
checkpoint['policy_optimizer_state_dict'])
if evaluate:
self.policy.eval()
self.critic.eval()
self.critic_target.eval()
else:
self.policy.train()
self.critic.train()
self.critic_target.train()