def input_prototype(self):
return rlt.PreprocessedState(
state=rlt.FeatureVector(
float_features=torch.randn(1, self.state_dim),
sequence_features=SequenceFeatures.prototype(),
)
)
def get_detached_q_values(
self, state
) -> Tuple[rlt.AllActionQValues, Optional[rlt.AllActionQValues]]:
""" Gets the q values from the model and target networks """
input = rlt.PreprocessedState(state=state)
q_values = self.q_network(input).q_values
q_values_target = self.q_network_target(input).q_values
return q_values, q_values_target
def input_prototype(self):
return rlt.PreprocessedState(state=rlt.PreprocessedFeatureVector(
float_features=torch.randn(1, self.state_dim),
id_list_features={
k: (
torch.zeros(1, dtype=torch.long),
torch.ones(1, dtype=torch.long),
)
for k in self.embedding_bags
},
))
def input_prototype(self):
return rlt.PreprocessedState(
state=rlt.FeatureVector(
float_features=torch.randn(1, self.state_dim),
id_list_features={
"page_id": (
torch.zeros(1, dtype=torch.long),
torch.ones(1, dtype=torch.long),
)
},
)
)
def internal_prediction(
self, state: torch.Tensor
) -> Union[rlt.SacPolicyActionSet, rlt.DqnPolicyActionSet]:
"""
Only used by Gym. Return the predicted next action
"""
input = rlt.PreprocessedState(state=rlt.PreprocessedFeatureVector(
float_features=state))
output = self.cem_planner_network(input)
if not self.cem_planner_network.discrete_action:
return rlt.SacPolicyActionSet(greedy=output, greedy_propensity=1.0)
return rlt.DqnPolicyActionSet(greedy=output[0])
def test_discrete_wrapper_with_id_list(self):
state_normalization_parameters = {i: _cont_norm() for i in range(1, 5)}
state_preprocessor = Preprocessor(state_normalization_parameters,
False)
action_dim = 2
state_feature_config = rlt.ModelFeatureConfig(
float_feature_infos=[
rlt.FloatFeatureInfo(name=str(i), feature_id=i)
for i in range(1, 5)
],
id_list_feature_configs=[
rlt.IdListFeatureConfig(name="A",
feature_id=10,
id_mapping_name="A_mapping")
],
id_mapping_config={"A_mapping": rlt.IdMapping(ids=[0, 1, 2])},
)
dqn = FullyConnectedDQNWithEmbedding(
state_dim=len(state_normalization_parameters),
action_dim=action_dim,
sizes=[16],
activations=["relu"],
model_feature_config=state_feature_config,
embedding_dim=8,
)
dqn_with_preprocessor = DiscreteDqnWithPreprocessorWithIdList(
dqn, state_preprocessor, state_feature_config)
action_names = ["L", "R"]
wrapper = DiscreteDqnPredictorWrapperWithIdList(
dqn_with_preprocessor, action_names, state_feature_config)
input_prototype = dqn_with_preprocessor.input_prototype()
output_action_names, q_values = wrapper(*input_prototype)
self.assertEqual(action_names, output_action_names)
self.assertEqual(q_values.shape, (1, 2))
feature_id_to_name = {
config.feature_id: config.name
for config in state_feature_config.id_list_feature_configs
}
state_id_list_features = {
feature_id_to_name[k]: v
for k, v in input_prototype[1].items()
}
expected_output = dqn(
rlt.PreprocessedState(state=rlt.PreprocessedFeatureVector(
float_features=state_preprocessor(*input_prototype[0]),
id_list_features=state_id_list_features,
))).q_values
self.assertTrue((expected_output == q_values).all())
def forward(
self,
state_with_presence: Tuple[torch.Tensor, torch.Tensor],
state_id_list_features: Dict[int, Tuple[torch.Tensor, torch.Tensor]],
):
preprocessed_state = self.state_preprocessor(state_with_presence[0],
state_with_presence[1])
id_list_features = {
id_list_feature_config.name:
state_id_list_features[id_list_feature_config.feature_id]
for id_list_feature_config in self.id_list_feature_configs
}
state_feature_vector = rlt.PreprocessedState(
state=rlt.PreprocessedFeatureVector(
float_features=preprocessed_state,
id_list_features=id_list_features))
q_values = self.model(state_feature_vector).q_values
return q_values
def train(self, training_batch):
if isinstance(training_batch, TrainingDataPage):
training_batch = training_batch.as_discrete_maxq_training_batch()
learning_input = training_batch.training_input
state = rlt.PreprocessedState(state=learning_input.state)
next_state = rlt.PreprocessedState(state=learning_input.next_state)
rewards = self.boost_rewards(learning_input.reward,
learning_input.action)
discount_tensor = torch.full_like(rewards, self.gamma)
possible_next_actions_mask = learning_input.possible_next_actions_mask.float(
)
possible_actions_mask = learning_input.possible_actions_mask.float()
self.minibatch += 1
not_done_mask = learning_input.not_terminal.float()
if self.use_seq_num_diff_as_time_diff:
assert self.multi_steps is None
discount_tensor = torch.pow(self.gamma,
learning_input.time_diff.float())
if self.multi_steps is not None:
discount_tensor = torch.pow(self.gamma,
learning_input.step.float())
next_qf = self.q_network_target.dist(next_state)
if self.maxq_learning:
# Select distribution corresponding to max valued action
next_q_values = (self.q_network(next_state)
if self.double_q_learning else next_qf.mean(2))
next_action = self.argmax_with_mask(next_q_values,
possible_next_actions_mask)
next_qf = next_qf[range(rewards.shape[0]), next_action.reshape(-1)]
else:
next_qf = (next_qf *
learning_input.next_action.unsqueeze(-1)).sum(1)
# Build target distribution
target_Q = rewards + discount_tensor * not_done_mask * next_qf
with torch.enable_grad():
current_qf = self.q_network.dist(state)
# for reporting only
all_q_values = current_qf.mean(2).detach()
current_qf = (current_qf *
learning_input.action.unsqueeze(-1)).sum(1)
# (batch, atoms) -> (atoms, batch, 1) -> (atoms, batch, atoms)
td = target_Q.t().unsqueeze(-1) - current_qf
loss = (self.huber(td) * (self.quantiles -
(td.detach() < 0).float()).abs()).mean()
loss.backward()
self._maybe_run_optimizer(self.q_network_optimizer,
self.minibatches_per_step)
# Use the soft update rule to update target network
self._maybe_soft_update(self.q_network, self.q_network_target,
self.tau, self.minibatches_per_step)
model_action_idxs = self.argmax_with_mask(
all_q_values,
possible_actions_mask
if self.maxq_learning else learning_input.action,
)
self.loss_reporter.report(
td_loss=loss,
logged_actions=learning_input.action.argmax(dim=1, keepdim=True),
logged_propensities=training_batch.extras.action_probability,
logged_rewards=rewards,
model_values=all_q_values,
model_action_idxs=model_action_idxs,
)
def create_from_tensors_dqn(
cls,
trainer: DQNTrainer,
mdp_ids: np.ndarray,
sequence_numbers: torch.Tensor,
states: rlt.PreprocessedFeatureVector,
actions: rlt.PreprocessedFeatureVector,
propensities: torch.Tensor,
rewards: torch.Tensor,
possible_actions_mask: torch.Tensor,
metrics: Optional[torch.Tensor] = None,
):
old_q_train_state = trainer.q_network.training
old_reward_train_state = trainer.reward_network.training
old_q_cpe_train_state = trainer.q_network_cpe.training
trainer.q_network.train(False)
trainer.reward_network.train(False)
trainer.q_network_cpe.train(False)
num_actions = trainer.num_actions
action_mask = actions.float() # type: ignore
rewards = trainer.boost_rewards(rewards, actions) # type: ignore
model_values = trainer.q_network_cpe(
rlt.PreprocessedState(state=states)
).q_values[:, 0:num_actions]
optimal_q_values, _ = trainer.get_detached_q_values(
states # type: ignore
)
eval_action_idxs = trainer.get_max_q_values( # type: ignore
optimal_q_values, possible_actions_mask
)[1]
model_propensities = masked_softmax(
optimal_q_values, possible_actions_mask, trainer.rl_temperature
)
assert model_values.shape == actions.shape, ( # type: ignore
"Invalid shape: "
+ str(model_values.shape) # type: ignore
+ " != "
+ str(actions.shape) # type: ignore
)
assert model_values.shape == possible_actions_mask.shape, ( # type: ignore
"Invalid shape: "
+ str(model_values.shape) # type: ignore
+ " != "
+ str(possible_actions_mask.shape) # type: ignore
)
model_values_for_logged_action = torch.sum(
model_values * action_mask, dim=1, keepdim=True
)
rewards_and_metric_rewards = trainer.reward_network(
rlt.PreprocessedState(state=states)
)
# In case we reuse the modular for Q-network
if hasattr(rewards_and_metric_rewards, "q_values"):
rewards_and_metric_rewards = rewards_and_metric_rewards.q_values
model_rewards = rewards_and_metric_rewards[:, 0:num_actions]
assert model_rewards.shape == actions.shape, ( # type: ignore
"Invalid shape: "
+ str(model_rewards.shape) # type: ignore
+ " != "
+ str(actions.shape) # type: ignore
)
model_rewards_for_logged_action = torch.sum(
model_rewards * action_mask, dim=1, keepdim=True
)
model_metrics = rewards_and_metric_rewards[:, num_actions:]
assert model_metrics.shape[1] % num_actions == 0, (
"Invalid metrics shape: "
+ str(model_metrics.shape)
+ " "
+ str(num_actions)
)
num_metrics = model_metrics.shape[1] // num_actions
if num_metrics == 0:
model_metrics_values = None
model_metrics_for_logged_action = None
model_metrics_values_for_logged_action = None
else:
model_metrics_values = trainer.q_network_cpe(
rlt.PreprocessedState(state=states)
)
# Backward compatility
if hasattr(model_metrics_values, "q_values"):
model_metrics_values = model_metrics_values.q_values
model_metrics_values = model_metrics_values[:, num_actions:]
assert (
model_metrics_values.shape[1] == num_actions * num_metrics
), ( # type: ignore
"Invalid shape: "
+ str(model_metrics_values.shape[1]) # type: ignore
+ " != "
+ str(actions.shape[1] * num_metrics) # type: ignore
)
model_metrics_for_logged_action_list = []
model_metrics_values_for_logged_action_list = []
for metric_index in range(num_metrics):
metric_start = metric_index * num_actions
metric_end = (metric_index + 1) * num_actions
model_metrics_for_logged_action_list.append(
torch.sum(
model_metrics[:, metric_start:metric_end] * action_mask,
dim=1,
keepdim=True,
)
)
model_metrics_values_for_logged_action_list.append(
torch.sum(
model_metrics_values[:, metric_start:metric_end] * action_mask,
dim=1,
keepdim=True,
)
)
model_metrics_for_logged_action = torch.cat(
model_metrics_for_logged_action_list, dim=1
)
model_metrics_values_for_logged_action = torch.cat(
model_metrics_values_for_logged_action_list, dim=1
)
trainer.q_network_cpe.train(old_q_cpe_train_state) # type: ignore
trainer.q_network.train(old_q_train_state) # type: ignore
trainer.reward_network.train(old_reward_train_state) # type: ignore
return cls(
mdp_id=mdp_ids,
sequence_number=sequence_numbers,
logged_propensities=propensities,
logged_rewards=rewards,
action_mask=action_mask,
model_rewards=model_rewards,
model_rewards_for_logged_action=model_rewards_for_logged_action,
model_values=model_values,
model_values_for_logged_action=model_values_for_logged_action,
model_metrics_values=model_metrics_values,
model_metrics_values_for_logged_action=model_metrics_values_for_logged_action,
model_propensities=model_propensities,
logged_metrics=metrics,
model_metrics=model_metrics,
model_metrics_for_logged_action=model_metrics_for_logged_action,
# Will compute later
logged_values=None,
logged_metrics_values=None,
possible_actions_mask=possible_actions_mask,
optimal_q_values=optimal_q_values,
eval_action_idxs=eval_action_idxs,
)
def train(self, training_batch):
if isinstance(training_batch, TrainingDataPage):
training_batch = training_batch.as_discrete_maxq_training_batch()
learning_input = training_batch.training_input
boosted_rewards = self.boost_rewards(learning_input.reward,
learning_input.action)
self.minibatch += 1
rewards = boosted_rewards
discount_tensor = torch.full_like(rewards, self.gamma)
not_done_mask = learning_input.not_terminal.float()
if self.use_seq_num_diff_as_time_diff:
assert self.multi_steps is None
discount_tensor = torch.pow(self.gamma,
learning_input.time_diff.float())
if self.multi_steps is not None:
discount_tensor = torch.pow(self.gamma,
learning_input.step.float())
all_next_q_values, all_next_q_values_target = self.get_detached_q_values(
learning_input.next_state)
if self.maxq_learning:
# Compute max a' Q(s', a') over all possible actions using target network
possible_next_actions_mask = (
learning_input.possible_next_actions_mask.float())
if self.bcq:
action_on_policy = get_valid_actions_from_imitator(
self.bcq_imitator,
learning_input.next_state,
self.bcq_drop_threshold,
)
possible_next_actions_mask *= action_on_policy
next_q_values, max_q_action_idxs = self.get_max_q_values_with_target(
all_next_q_values, all_next_q_values_target,
possible_next_actions_mask)
else:
# SARSA
next_q_values, max_q_action_idxs = self.get_max_q_values_with_target(
all_next_q_values, all_next_q_values_target,
learning_input.next_action)
filtered_next_q_vals = next_q_values * not_done_mask
target_q_values = rewards + (discount_tensor * filtered_next_q_vals)
with torch.enable_grad():
# Get Q-value of action taken
current_state = rlt.PreprocessedState(state=learning_input.state)
all_q_values = self.q_network(current_state).q_values
self.all_action_scores = all_q_values.detach()
q_values = torch.sum(all_q_values * learning_input.action,
1,
keepdim=True)
loss = self.q_network_loss(q_values, target_q_values)
self.loss = loss.detach()
loss.backward()
self._maybe_run_optimizer(self.q_network_optimizer,
self.minibatches_per_step)
# Use the soft update rule to update target network
self._maybe_soft_update(self.q_network, self.q_network_target,
self.tau, self.minibatches_per_step)
# Get Q-values of next states, used in computing cpe
next_state = rlt.PreprocessedState(state=learning_input.next_state)
all_next_action_scores = self.q_network(next_state).q_values.detach()
logged_action_idxs = learning_input.action.argmax(dim=1, keepdim=True)
reward_loss, model_rewards, model_propensities = self._calculate_cpes(
training_batch,
current_state,
next_state,
self.all_action_scores,
all_next_action_scores,
logged_action_idxs,
discount_tensor,
not_done_mask,
)
if self.maxq_learning:
possible_actions_mask = learning_input.possible_actions_mask
if self.bcq:
action_on_policy = get_valid_actions_from_imitator(
self.bcq_imitator, learning_input.state,
self.bcq_drop_threshold)
possible_actions_mask *= action_on_policy
model_action_idxs = self.get_max_q_values(
self.all_action_scores,
possible_actions_mask
if self.maxq_learning else learning_input.action,
)[1]
self.loss_reporter.report(
td_loss=self.loss,
reward_loss=reward_loss,
logged_actions=logged_action_idxs,
logged_propensities=training_batch.extras.action_probability,
logged_rewards=rewards,
logged_values=None, # Compute at end of each epoch for CPE
model_propensities=model_propensities,
model_rewards=model_rewards,
model_values=self.all_action_scores,
model_values_on_logged_actions=
None, # Compute at end of each epoch for CPE
model_action_idxs=model_action_idxs,
)
def train(self, training_batch):
if isinstance(training_batch, TrainingDataPage):
training_batch = training_batch.as_discrete_maxq_training_batch()
learning_input = training_batch.training_input
state = rlt.PreprocessedState(state=learning_input.state)
next_state = rlt.PreprocessedState(state=learning_input.next_state)
rewards = self.boost_rewards(learning_input.reward,
learning_input.action)
discount_tensor = torch.full_like(rewards, self.gamma)
possible_next_actions_mask = learning_input.possible_next_actions_mask.float(
)
possible_actions_mask = learning_input.possible_actions_mask.float()
self.minibatch += 1
not_done_mask = learning_input.not_terminal.float()
if self.use_seq_num_diff_as_time_diff:
assert self.multi_steps is None
discount_tensor = torch.pow(self.gamma,
learning_input.time_diff.float())
if self.multi_steps is not None:
discount_tensor = torch.pow(self.gamma,
learning_input.step.float())
next_dist = self.q_network_target.log_dist(next_state).exp()
if self.maxq_learning:
# Select distribution corresponding to max valued action
if self.double_q_learning:
next_q_values = (self.q_network.log_dist(next_state).exp() *
self.support).sum(2)
else:
next_q_values = (next_dist * self.support).sum(2)
next_action = self.argmax_with_mask(next_q_values,
possible_next_actions_mask)
next_dist = next_dist[range(rewards.shape[0]),
next_action.reshape(-1)]
else:
next_dist = (next_dist *
learning_input.next_action.unsqueeze(-1)).sum(1)
# Build target distribution
target_Q = rewards + discount_tensor * not_done_mask * self.support
# Project target distribution back onto support
# remove support outliers
target_Q = target_Q.clamp(self.qmin, self.qmax)
# rescale to indicies
b = (target_Q - self.qmin) / (self.qmax -
self.qmin) * (self.num_atoms - 1.0)
lower = b.floor()
upper = b.ceil()
# Since index_add_ doesn't work with multiple dimensions
# we operate on the flattened tensors
offset = self.num_atoms * torch.arange(
rewards.shape[0], device=self.device, dtype=torch.long).reshape(
-1, 1).repeat(1, self.num_atoms)
m = torch.zeros_like(next_dist)
m.reshape(-1).index_add_(
0,
(lower.long() + offset).reshape(-1),
(next_dist * (upper - b)).reshape(-1),
)
m.reshape(-1).index_add_(
0,
(upper.long() + offset).reshape(-1),
(next_dist * (b - lower)).reshape(-1),
)
with torch.enable_grad():
log_dist = self.q_network.log_dist(state)
# for reporting only
all_q_values = (log_dist.exp() * self.support).sum(2).detach()
log_dist = (log_dist * learning_input.action.unsqueeze(-1)).sum(1)
loss = -(m * log_dist).sum(1).mean()
loss.backward()
self._maybe_run_optimizer(self.q_network_optimizer,
self.minibatches_per_step)
# Use the soft update rule to update target network
self._maybe_soft_update(self.q_network, self.q_network_target,
self.tau, self.minibatches_per_step)
model_action_idxs = self.argmax_with_mask(
all_q_values,
possible_actions_mask
if self.maxq_learning else learning_input.action,
)
self.loss_reporter.report(
td_loss=loss,
logged_actions=learning_input.action.argmax(dim=1, keepdim=True),
logged_propensities=training_batch.extras.action_probability,
logged_rewards=rewards,
model_values=all_q_values,
model_action_idxs=model_action_idxs,
)
def get_detached_q_values(self, state):
""" Gets the q values from the model and target networks """
input = rlt.PreprocessedState(state=state)
q_values = self.q_network(input).q_values
q_values_target = self.q_network_target(input).q_values
return q_values, q_values_target
def train(self, training_batch) -> None:
"""
IMPORTANT: the input action here is assumed to be preprocessed to match the
range of the output of the actor.
"""
if hasattr(training_batch, "as_policy_network_training_batch"):
training_batch = training_batch.as_policy_network_training_batch()
learning_input = training_batch.training_input
self.minibatch += 1
state = learning_input.state
action = learning_input.action
next_state = learning_input.next_state
reward = learning_input.reward
not_done_mask = learning_input.not_terminal
action = self._maybe_scale_action_in_train(action.float_features)
max_action = (self.max_action_range_tensor_training
if self.max_action_range_tensor_training else torch.ones(
action.shape, device=self.device))
min_action = (self.min_action_range_tensor_serving
if self.min_action_range_tensor_serving else
-torch.ones(action.shape, device=self.device))
# Compute current value estimates
current_state_action = rlt.PreprocessedStateAction(
state=state,
action=rlt.PreprocessedFeatureVector(float_features=action))
q1_value = self.q1_network(current_state_action).q_value
if self.q2_network:
q2_value = self.q2_network(current_state_action).q_value
actor_action = self.actor_network(
rlt.PreprocessedState(state=state)).action
# Generate target = r + y * min (Q1(s',pi(s')), Q2(s',pi(s')))
with torch.no_grad():
next_actor = self.actor_network_target(
rlt.PreprocessedState(state=next_state)).action
next_actor += (torch.randn_like(next_actor) *
self.target_policy_smoothing).clamp(
-self.noise_clip, self.noise_clip)
next_actor = torch.max(torch.min(next_actor, max_action),
min_action)
next_state_actor = rlt.PreprocessedStateAction(
state=next_state,
action=rlt.PreprocessedFeatureVector(
float_features=next_actor),
)
next_state_value = self.q1_network_target(next_state_actor).q_value
if self.q2_network is not None:
next_state_value = torch.min(
next_state_value,
self.q2_network_target(next_state_actor).q_value)
target_q_value = (
reward + self.gamma * next_state_value * not_done_mask.float())
# Optimize Q1 and Q2
q1_loss = F.mse_loss(q1_value, target_q_value)
q1_loss.backward()
self._maybe_run_optimizer(self.q1_network_optimizer,
self.minibatches_per_step)
if self.q2_network:
q2_loss = F.mse_loss(q2_value, target_q_value)
q2_loss.backward()
self._maybe_run_optimizer(self.q2_network_optimizer,
self.minibatches_per_step)
# Only update actor and target networks after a fixed number of Q updates
if self.minibatch % self.delayed_policy_update == 0:
actor_loss = -self.q1_network(
rlt.PreprocessedStateAction(
state=state,
action=rlt.PreprocessedFeatureVector(
float_features=actor_action),
)).q_value.mean()
actor_loss.backward()
self._maybe_run_optimizer(self.actor_network_optimizer,
self.minibatches_per_step)
# Use the soft update rule to update the target networks
self._maybe_soft_update(
self.q1_network,
self.q1_network_target,
self.tau,
self.minibatches_per_step,
)
self._maybe_soft_update(
self.actor_network,
self.actor_network_target,
self.tau,
self.minibatches_per_step,
)
if self.q2_network is not None:
self._maybe_soft_update(
self.q2_network,
self.q2_network_target,
self.tau,
self.minibatches_per_step,
)
# Logging at the end to schedule all the cuda operations first
if (self.tensorboard_logging_freq != 0
and self.minibatch % self.tensorboard_logging_freq == 0):
SummaryWriterContext.add_histogram("q1/logged_state_value",
q1_value)
if self.q2_network:
SummaryWriterContext.add_histogram("q2/logged_state_value",
q2_value)
SummaryWriterContext.add_histogram("q_network/next_state_value",
next_state_value)
SummaryWriterContext.add_histogram("q_network/target_q_value",
target_q_value)
SummaryWriterContext.add_histogram("actor/loss", actor_loss)
self.loss_reporter.report(
td_loss=float(q1_loss),
reward_loss=None,
logged_rewards=reward,
model_values_on_logged_actions=q1_value,
)
def create_from_tensors(
cls,
trainer: DQNTrainer,
mdp_ids: np.ndarray,
sequence_numbers: torch.Tensor,
states: rlt.PreprocessedFeatureVector,
actions: rlt.PreprocessedFeatureVector,
propensities: torch.Tensor,
rewards: torch.Tensor,
possible_actions_mask: torch.Tensor,
possible_actions: Optional[rlt.PreprocessedFeatureVector] = None,
max_num_actions: Optional[int] = None,
metrics: Optional[torch.Tensor] = None,
):
# Switch to evaluation mode for the network
old_q_train_state = trainer.q_network.training
old_reward_train_state = trainer.reward_network.training
trainer.q_network.train(False)
trainer.reward_network.train(False)
if max_num_actions:
# Parametric model CPE
state_action_pairs = rlt.PreprocessedStateAction(
state=states, action=actions
)
tiled_state = states.float_features.repeat(1, max_num_actions).reshape(
-1, states.float_features.shape[1]
)
assert possible_actions is not None
# Get Q-value of action taken
possible_actions_state_concat = rlt.PreprocessedStateAction(
state=rlt.PreprocessedFeatureVector(float_features=tiled_state),
action=possible_actions,
)
# Parametric actions
# FIXME: model_values and model propensities should be calculated
# as in discrete dqn model
model_values = trainer.q_network(
possible_actions_state_concat
).q_value # type: ignore
optimal_q_values = model_values
eval_action_idxs = None
assert (
model_values.shape[0] * model_values.shape[1]
== possible_actions_mask.shape[0] * possible_actions_mask.shape[1]
), (
"Invalid shapes: "
+ str(model_values.shape)
+ " != "
+ str(possible_actions_mask.shape)
)
model_values = model_values.reshape(possible_actions_mask.shape)
model_propensities = masked_softmax(
model_values, possible_actions_mask, trainer.rl_temperature
)
model_rewards = trainer.reward_network(
possible_actions_state_concat
).q_value # type: ignore
assert (
model_rewards.shape[0] * model_rewards.shape[1]
== possible_actions_mask.shape[0] * possible_actions_mask.shape[1]
), (
"Invalid shapes: "
+ str(model_rewards.shape)
+ " != "
+ str(possible_actions_mask.shape)
)
model_rewards = model_rewards.reshape(possible_actions_mask.shape)
model_values_for_logged_action = trainer.q_network(
state_action_pairs
).q_value
model_rewards_for_logged_action = trainer.reward_network(
state_action_pairs
).q_value
action_mask = (
torch.abs(model_values - model_values_for_logged_action) < 1e-3
).float()
model_metrics = None
model_metrics_for_logged_action = None
model_metrics_values = None
model_metrics_values_for_logged_action = None
else:
num_actions = trainer.num_actions
action_mask = actions.float() # type: ignore
# Switch to evaluation mode for the network
old_q_cpe_train_state = trainer.q_network_cpe.training
trainer.q_network_cpe.train(False)
# Discrete actions
rewards = trainer.boost_rewards(rewards, actions) # type: ignore
model_values = trainer.q_network_cpe(
rlt.PreprocessedState(state=states)
).q_values[:, 0:num_actions]
optimal_q_values = trainer.get_detached_q_values(
states # type: ignore
)[ # type: ignore
0
] # type: ignore
eval_action_idxs = trainer.get_max_q_values( # type: ignore
optimal_q_values, possible_actions_mask
)[1]
model_propensities = masked_softmax(
optimal_q_values, possible_actions_mask, trainer.rl_temperature
)
assert model_values.shape == actions.shape, ( # type: ignore
"Invalid shape: "
+ str(model_values.shape) # type: ignore
+ " != "
+ str(actions.shape) # type: ignore
)
assert model_values.shape == possible_actions_mask.shape, ( # type: ignore
"Invalid shape: "
+ str(model_values.shape) # type: ignore
+ " != "
+ str(possible_actions_mask.shape) # type: ignore
)
model_values_for_logged_action = torch.sum(
model_values * action_mask, dim=1, keepdim=True
)
rewards_and_metric_rewards = trainer.reward_network(
rlt.PreprocessedState(state=states)
)
# In case we reuse the modular for Q-network
if hasattr(rewards_and_metric_rewards, "q_values"):
rewards_and_metric_rewards = rewards_and_metric_rewards.q_values
model_rewards = rewards_and_metric_rewards[:, 0:num_actions]
assert model_rewards.shape == actions.shape, ( # type: ignore
"Invalid shape: "
+ str(model_rewards.shape) # type: ignore
+ " != "
+ str(actions.shape) # type: ignore
)
model_rewards_for_logged_action = torch.sum(
model_rewards * action_mask, dim=1, keepdim=True
)
model_metrics = rewards_and_metric_rewards[:, num_actions:]
assert model_metrics.shape[1] % num_actions == 0, (
"Invalid metrics shape: "
+ str(model_metrics.shape)
+ " "
+ str(num_actions)
)
num_metrics = model_metrics.shape[1] // num_actions
if num_metrics == 0:
model_metrics_values = None
model_metrics_for_logged_action = None
model_metrics_values_for_logged_action = None
else:
model_metrics_values = trainer.q_network_cpe(
rlt.PreprocessedState(state=states)
)
# Backward compatility
if hasattr(model_metrics_values, "q_values"):
model_metrics_values = model_metrics_values.q_values
model_metrics_values = model_metrics_values[:, num_actions:]
assert (
model_metrics_values.shape[1] == num_actions * num_metrics
), ( # type: ignore
"Invalid shape: "
+ str(model_metrics_values.shape[1]) # type: ignore
+ " != "
+ str(actions.shape[1] * num_metrics) # type: ignore
)
model_metrics_for_logged_action_list = []
model_metrics_values_for_logged_action_list = []
for metric_index in range(num_metrics):
metric_start = metric_index * num_actions
metric_end = (metric_index + 1) * num_actions
model_metrics_for_logged_action_list.append(
torch.sum(
model_metrics[:, metric_start:metric_end] * action_mask,
dim=1,
keepdim=True,
)
)
model_metrics_values_for_logged_action_list.append(
torch.sum(
model_metrics_values[:, metric_start:metric_end]
* action_mask,
dim=1,
keepdim=True,
)
)
model_metrics_for_logged_action = torch.cat(
model_metrics_for_logged_action_list, dim=1
)
model_metrics_values_for_logged_action = torch.cat(
model_metrics_values_for_logged_action_list, dim=1
)
# Switch back to the old mode
trainer.q_network_cpe.train(old_q_cpe_train_state) # type: ignore
# Switch back to the old mode
trainer.q_network.train(old_q_train_state) # type: ignore
trainer.reward_network.train(old_reward_train_state) # type: ignore
return cls(
mdp_id=mdp_ids,
sequence_number=sequence_numbers,
logged_propensities=propensities,
logged_rewards=rewards,
action_mask=action_mask,
model_rewards=model_rewards,
model_rewards_for_logged_action=model_rewards_for_logged_action,
model_values=model_values,
model_values_for_logged_action=model_values_for_logged_action,
model_metrics_values=model_metrics_values,
model_metrics_values_for_logged_action=model_metrics_values_for_logged_action,
model_propensities=model_propensities,
logged_metrics=metrics,
model_metrics=model_metrics,
model_metrics_for_logged_action=model_metrics_for_logged_action,
# Will compute later
logged_values=None,
logged_metrics_values=None,
possible_actions_mask=possible_actions_mask,
optimal_q_values=optimal_q_values,
eval_action_idxs=eval_action_idxs,
)
def train(self, training_batch) -> None:
"""
IMPORTANT: the input action here is assumed to be preprocessed to match the
range of the output of the actor.
"""
if hasattr(training_batch, "as_policy_network_training_batch"):
training_batch = training_batch.as_policy_network_training_batch()
learning_input = training_batch.training_input
self.minibatch += 1
state = learning_input.state
action = learning_input.action
reward = learning_input.reward
discount = torch.full_like(reward, self.gamma)
not_done_mask = learning_input.not_terminal
if self._should_scale_action_in_train():
action = action._replace(
float_features=rescale_torch_tensor(
action.float_features,
new_min=self.min_action_range_tensor_training,
new_max=self.max_action_range_tensor_training,
prev_min=self.min_action_range_tensor_serving,
prev_max=self.max_action_range_tensor_serving,
)
)
with torch.enable_grad():
#
# First, optimize Q networks; minimizing MSE between
# Q(s, a) & r + discount * V'(next_s)
#
current_state_action = rlt.PreprocessedStateAction(
state=state, action=action
)
q1_value = self.q1_network(current_state_action).q_value
if self.q2_network:
q2_value = self.q2_network(current_state_action).q_value
actor_output = self.actor_network(rlt.PreprocessedState(state=state))
# Optimize Alpha
if self.alpha_optimizer is not None:
alpha_loss = -(
self.log_alpha
* (actor_output.log_prob + self.target_entropy).detach()
).mean()
self.alpha_optimizer.zero_grad()
alpha_loss.backward()
self.alpha_optimizer.step()
self.entropy_temperature = self.log_alpha.exp()
with torch.no_grad():
if self.value_network is not None:
next_state_value = self.value_network_target(
learning_input.next_state.float_features
)
else:
next_state_actor_output = self.actor_network(
rlt.PreprocessedState(state=learning_input.next_state)
)
next_state_actor_action = rlt.PreprocessedStateAction(
state=learning_input.next_state,
action=rlt.PreprocessedFeatureVector(
float_features=next_state_actor_output.action
),
)
next_state_value = self.q1_network_target(
next_state_actor_action
).q_value
if self.q2_network is not None:
target_q2_value = self.q2_network_target(
next_state_actor_action
).q_value
next_state_value = torch.min(next_state_value, target_q2_value)
log_prob_a = self.actor_network.get_log_prob(
learning_input.next_state, next_state_actor_output.action
)
log_prob_a = log_prob_a.clamp(-20.0, 20.0)
next_state_value -= self.entropy_temperature * log_prob_a
target_q_value = (
reward + discount * next_state_value * not_done_mask.float()
)
q1_loss = F.mse_loss(q1_value, target_q_value)
q1_loss.backward()
self._maybe_run_optimizer(
self.q1_network_optimizer, self.minibatches_per_step
)
if self.q2_network:
q2_loss = F.mse_loss(q2_value, target_q_value)
q2_loss.backward()
self._maybe_run_optimizer(
self.q2_network_optimizer, self.minibatches_per_step
)
#
# Second, optimize the actor; minimizing KL-divergence between action propensity
# & softmax of value. Due to reparameterization trick, it ends up being
# log_prob(actor_action) - Q(s, actor_action)
#
state_actor_action = rlt.PreprocessedStateAction(
state=state,
action=rlt.PreprocessedFeatureVector(
float_features=actor_output.action
),
)
q1_actor_value = self.q1_network(state_actor_action).q_value
min_q_actor_value = q1_actor_value
if self.q2_network:
q2_actor_value = self.q2_network(state_actor_action).q_value
min_q_actor_value = torch.min(q1_actor_value, q2_actor_value)
actor_loss = (
self.entropy_temperature * actor_output.log_prob - min_q_actor_value
)
# Do this in 2 steps so we can log histogram of actor loss
actor_loss_mean = actor_loss.mean()
actor_loss_mean.backward()
self._maybe_run_optimizer(
self.actor_network_optimizer, self.minibatches_per_step
)
#
# Lastly, if applicable, optimize value network; minimizing MSE between
# V(s) & E_a~pi(s) [ Q(s,a) - log(pi(a|s)) ]
#
if self.value_network is not None:
state_value = self.value_network(state.float_features)
if self.logged_action_uniform_prior:
log_prob_a = torch.zeros_like(min_q_actor_value)
target_value = min_q_actor_value
else:
with torch.no_grad():
log_prob_a = actor_output.log_prob
log_prob_a = log_prob_a.clamp(-20.0, 20.0)
target_value = (
min_q_actor_value - self.entropy_temperature * log_prob_a
)
value_loss = F.mse_loss(state_value, target_value.detach())
value_loss.backward()
self._maybe_run_optimizer(
self.value_network_optimizer, self.minibatches_per_step
)
# Use the soft update rule to update the target networks
if self.value_network is not None:
self._maybe_soft_update(
self.value_network,
self.value_network_target,
self.tau,
self.minibatches_per_step,
)
else:
self._maybe_soft_update(
self.q1_network,
self.q1_network_target,
self.tau,
self.minibatches_per_step,
)
if self.q2_network is not None:
self._maybe_soft_update(
self.q2_network,
self.q2_network_target,
self.tau,
self.minibatches_per_step,
)
# Logging at the end to schedule all the cuda operations first
if (
self.tensorboard_logging_freq is not None
and self.minibatch % self.tensorboard_logging_freq == 0
):
SummaryWriterContext.add_histogram("q1/logged_state_value", q1_value)
if self.q2_network:
SummaryWriterContext.add_histogram("q2/logged_state_value", q2_value)
SummaryWriterContext.add_histogram("log_prob_a", log_prob_a)
if self.value_network:
SummaryWriterContext.add_histogram("value_network/target", target_value)
SummaryWriterContext.add_histogram(
"q_network/next_state_value", next_state_value
)
SummaryWriterContext.add_histogram(
"q_network/target_q_value", target_q_value
)
SummaryWriterContext.add_histogram(
"actor/min_q_actor_value", min_q_actor_value
)
SummaryWriterContext.add_histogram(
"actor/action_log_prob", actor_output.log_prob
)
SummaryWriterContext.add_histogram("actor/loss", actor_loss)
self.loss_reporter.report(
td_loss=float(q1_loss),
reward_loss=None,
logged_rewards=reward,
model_values_on_logged_actions=q1_value,
model_propensities=actor_output.log_prob.exp(),
model_values=min_q_actor_value,
)