# -*- coding: utf-8 -*-
"""Training loops for KGE models using multi-modal information."""
import gc
import logging
from abc import ABC, abstractmethod
from typing import Any, List, Mapping, Optional, Tuple, Type, Union
import torch
from torch.optim.optimizer import Optimizer
from torch.utils.data import DataLoader
from tqdm import tqdm, trange
from ..losses import Loss
from ..models.base import Model
from ..stoppers import Stopper
from ..training.schlichtkrull_sampler import GraphSampler
from ..triples import Instances, TriplesFactory
from ..typing import MappedTriples
from ..utils import ResultTracker, is_cuda_oom_error, is_cudnn_error, normalize_string
__all__ = [
'TrainingLoop',
'NonFiniteLossError',
'TrainingApproachLossMismatchError',
'SubBatchingNotSupportedError',
]
logger = logging.getLogger(__name__)
class NonFiniteLossError(RuntimeError):
"""An exception raised for non-finite loss values."""
class TrainingApproachLossMismatchError(TypeError):
"""An exception when an illegal loss function is used with a given training approach."""
class SubBatchingNotSupportedError(NotImplementedError):
"""An exception raised when sub batching is not implemented."""
def __init__(self, model: Model):
super().__init__(model)
self.model = model
def __str__(self): # noqa: D105
return f'No sub-batching support for {self.model.__class__.__name__} due to modules ' \
f'{self.model.modules_not_supporting_sub_batching}.'
def _get_optimizer_kwargs(optimizer: Optimizer) -> Mapping[str, Any]:
optimizer_kwargs = optimizer.state_dict()
optimizer_kwargs = {
key: value
for key, value in optimizer_kwargs['param_groups'][0].items()
if key != 'params'
}
return optimizer_kwargs
class TrainingLoop(ABC):
"""A training loop."""
training_instances: Optional[Instances]
losses_per_epochs: List[float]
loss_blacklist: Optional[List[Type[Loss]]] = None
hpo_default = dict(
num_epochs=dict(type=int, low=100, high=1000, q=100),
batch_size=dict(type=int, low=32, high=4000, q=100),
)
def __init__(
self,
model: Model,
optimizer: Optional[Optimizer] = None,
) -> None:
"""Initialize the training loop.
:param model: The model to train
:param optimizer: The optimizer to use while training the model
"""
self.model = model
self.optimizer = optimizer
self.training_instances = None
self.losses_per_epochs = []
if self.loss_blacklist and isinstance(self.model.loss, tuple(self.loss_blacklist)):
raise TrainingApproachLossMismatchError(
f'Can not use loss {self.model.loss.__class__.__name__}'
f' with training approach {self.__class__.__name__}',
)
if self.model.is_mr_loss:
self._loss_helper = self._mr_loss_helper
elif self.model.is_nssa_loss:
self._loss_helper = self._self_adversarial_negative_sampling_loss_helper
else:
self._loss_helper = self._label_loss_helper
@classmethod
def get_normalized_name(cls) -> str:
"""Get the normalized name of the training loop."""
return normalize_string(cls.__name__, suffix=TrainingLoop.__name__)
@property
def triples_factory(self) -> TriplesFactory: # noqa: D401
"""The triples factory in the model."""
return self.model.triples_factory
@property
def device(self): # noqa: D401
"""The device used by the model."""
return self.model.device
def train(
self,
num_epochs: int = 1,
batch_size: Optional[int] = None,
slice_size: Optional[int] = None,
label_smoothing: float = 0.0,
sampler: Optional[str] = None,
continue_training: bool = False,
only_size_probing: bool = False,
tqdm_kwargs: Optional[Mapping[str, Any]] = None,
stopper: Optional[Stopper] = None,
result_tracker: Optional[ResultTracker] = None,
sub_batch_size: Optional[int] = None,
num_workers: Optional[int] = None,
clear_optimizer: bool = False,
) -> List[float]:
"""Train the KGE model.
:param num_epochs:
The number of epochs to train the model.
:param batch_size:
If set the batch size to use for mini-batch training. Otherwise find the largest possible batch_size
automatically.
:param slice_size: >0
The divisor for the scoring function when using slicing. This is only possible for LCWA training loops in
general and only for models that have the slicing capability implemented.
:param label_smoothing: (0 <= label_smoothing < 1)
If larger than zero, use label smoothing.
:param sampler: (None or 'schlichtkrull')
The type of sampler to use. At the moment sLCWA in R-GCN is the only user of schlichtkrull sampling.
:param continue_training:
If set to False, (re-)initialize the model's weights. Otherwise continue training.
:param only_size_probing:
The evaluation is only performed for two batches to test the memory footprint, especially on GPUs.
:param tqdm_kwargs:
Keyword arguments passed to :mod:`tqdm` managing the progress bar.
:param stopper:
An instance of :class:`pykeen.stopper.EarlyStopper` with settings for checking
if training should stop early
:param result_tracker:
The result tracker.
:param sub_batch_size:
If provided split each batch into sub-batches to avoid memory issues for large models / small GPUs.
:param num_workers:
The number of child CPU workers used for loading data. If None, data are loaded in the main process.
:param clear_optimizer:
Whether to delete the optimizer instance after training (as the optimizer might have additional memory
consumption due to e.g. moments in Adam).
:return:
A pair of the KGE model and the losses per epoch.
"""
# Create training instances
# During size probing the training instances should not show the tqdm progress bar
self.training_instances = self._create_instances(use_tqdm=not only_size_probing)
# In some cases, e.g. using Optuna for HPO, the cuda cache from a previous run is not cleared
torch.cuda.empty_cache()
result = self._train(
num_epochs=num_epochs,
batch_size=batch_size,
slice_size=slice_size,
label_smoothing=label_smoothing,
sampler=sampler,
continue_training=continue_training,
only_size_probing=only_size_probing,
tqdm_kwargs=tqdm_kwargs,
stopper=stopper,
result_tracker=result_tracker,
sub_batch_size=sub_batch_size,
num_workers=num_workers,
)
# Ensure the release of memory
torch.cuda.empty_cache()
# Clear optimizer
if clear_optimizer:
self.optimizer = None
return result
def _train( # noqa: C901
self,
num_epochs: int = 1,
batch_size: Optional[int] = None,
slice_size: Optional[int] = None,
label_smoothing: float = 0.0,
sampler: Optional[str] = None,
continue_training: bool = False,
only_size_probing: bool = False,
tqdm_kwargs: Optional[Mapping[str, Any]] = None,
stopper: Optional[Stopper] = None,
result_tracker: Optional[ResultTracker] = None,
sub_batch_size: Optional[int] = None,
num_workers: Optional[int] = None,
) -> List[float]:
"""Train the KGE model.
:param num_epochs:
The number of epochs to train the model.
:param batch_size:
If set the batch size to use for mini-batch training. Otherwise find the largest possible batch_size
automatically.
:param slice_size: >0
The divisor for the scoring function when using slicing. This is only possible for LCWA training loops in
general and only for models that have the slicing capability implemented.
:param label_smoothing: (0 <= label_smoothing < 1)
If larger than zero, use label smoothing.
:param sampler: (None or 'schlichtkrull')
The type of sampler to use. At the moment sLCWA in R-GCN is the only user of schlichtkrull sampling.
:param continue_training:
If set to False, (re-)initialize the model's weights. Otherwise continue training.
:param only_size_probing:
The evaluation is only performed for two batches to test the memory footprint, especially on GPUs.
:param tqdm_kwargs:
Keyword arguments passed to :mod:`tqdm` managing the progress bar.
:param stopper:
An instance of :class:`pykeen.stopper.Stopper` with settings for checking
if training should stop early
:param result_tracker:
The result tracker.
:param sub_batch_size:
If provided split each batch into sub-batches to avoid memory issues for large models / small GPUs.
:param num_workers:
The number of child CPU workers used for loading data. If None, data are loaded in the main process.
:return:
A pair of the KGE model and the losses per epoch.
"""
# Take the biggest possible training batch_size, if batch_size not set
batch_size_sufficient = False
if batch_size is None:
if self.model.automatic_memory_optimization:
batch_size, batch_size_sufficient = self.batch_size_search()
else:
batch_size = 256
# This will find necessary parameters to optimize the use of the hardware at hand
if not only_size_probing and self.model.automatic_memory_optimization and not batch_size_sufficient:
# return the relevant parameters slice_size and batch_size
sub_batch_size, slice_size = self.sub_batch_and_slice(batch_size)
# Create dummy result tracker
if result_tracker is None:
result_tracker = ResultTracker()
if sub_batch_size is None or sub_batch_size == batch_size: # by default do not split batches in sub-batches
sub_batch_size = batch_size
elif not self.model.supports_subbatching:
raise SubBatchingNotSupportedError(self.model)
# Sanity check
if self.model.is_mr_loss and label_smoothing > 0.:
raise RuntimeError('Label smoothing can not be used with margin ranking loss.')
# Force weight initialization if training continuation is not explicitly requested.
if not continue_training:
# Reset the weights
self.model.reset_parameters_()
# Create new optimizer
optimizer_kwargs = _get_optimizer_kwargs(self.optimizer)
self.optimizer = self.optimizer.__class__(
params=self.model.get_grad_params(),
**optimizer_kwargs
)
elif not self.optimizer.state:
raise ValueError('Cannot continue_training without being trained once.')
# Ensure the model is on the correct device
self.model: Model = self.model.to(self.device)
# Create Sampler
if sampler == 'schlichtkrull':
sampler = GraphSampler(self.triples_factory, num_samples=sub_batch_size)
shuffle = False
else:
sampler = None
shuffle = True
if num_workers is None:
num_workers = 0
# Bind
num_training_instances = self.training_instances.num_instances
# When size probing, we don't want progress bars
if not only_size_probing:
# Create progress bar
_tqdm_kwargs = dict(desc=f'Training epochs on {self.device}', unit='epoch')
if tqdm_kwargs is not None:
_tqdm_kwargs.update(tqdm_kwargs)
epochs = trange(1, 1 + num_epochs, **_tqdm_kwargs)
logger.info(f'using stopper: {stopper}')
else:
epochs = range(1, 1 + num_epochs)
logger.debug(f'using stopper: {stopper}')
train_data_loader = DataLoader(
sampler=sampler,
dataset=self.training_instances,
batch_size=batch_size,
shuffle=shuffle,
num_workers=num_workers,
)
# Training Loop
for epoch in epochs:
# Enforce training mode
self.model.train()
# Accumulate loss over epoch
current_epoch_loss = 0.
# Batching
# Only create a progress bar when not in size probing mode
if not only_size_probing:
batches = tqdm(train_data_loader, desc=f'Training batches on {self.device}', leave=False, unit='batch')
else:
batches = train_data_loader
# Flag to check when to quit the size probing
evaluated_once = False
for batch in batches:
# Recall that torch *accumulates* gradients. Before passing in a
# new instance, you need to zero out the gradients from the old instance
self.optimizer.zero_grad()
# Get batch size of current batch (last batch may be incomplete)
current_batch_size = self._get_batch_size(batch)
# accumulate gradients for whole batch
for start in range(0, current_batch_size, sub_batch_size):
stop = min(start + sub_batch_size, current_batch_size)
# forward pass call
current_epoch_loss += self._forward_pass(
batch,
start,
stop,
current_batch_size,
label_smoothing,
slice_size,
)
# when called by batch_size_search(), the parameter update should not be applied.
if not only_size_probing:
# update parameters according to optimizer
self.optimizer.step()
# After changing applying the gradients to the embeddings, the model is notified that the forward
# constraints are no longer applied
self.model.post_parameter_update()
# For testing purposes we're only interested in processing one batch
if only_size_probing and evaluated_once:
break
evaluated_once = True
del batch
del batches
gc.collect()
self.optimizer.zero_grad()
self._free_graph_and_cache()
# When size probing we don't need the losses
if only_size_probing:
return None
# Track epoch loss
epoch_loss = current_epoch_loss / num_training_instances
self.losses_per_epochs.append(epoch_loss)
result_tracker.log_metrics({'loss': epoch_loss}, step=epoch)
# Print loss information to console
epochs.set_postfix({
'loss': self.losses_per_epochs[-1],
'prev_loss': self.losses_per_epochs[-2] if epoch > 2 else float('nan'),
})
if stopper is not None and stopper.should_evaluate(epoch) and stopper.should_stop():
return self.losses_per_epochs
return self.losses_per_epochs
def _forward_pass(self, batch, start, stop, current_batch_size, label_smoothing, slice_size):
# forward pass
loss = self._process_batch(
batch=batch,
start=start,
stop=stop,
label_smoothing=label_smoothing,
slice_size=slice_size
)
# raise error when non-finite loss occurs (NaN, +/-inf)
if not torch.isfinite(loss):
raise NonFiniteLossError('Loss is non-finite.')
# correction for loss reduction
if self.model.loss.reduction == 'mean':
this_sub_batch_size = stop - start
loss *= (this_sub_batch_size / current_batch_size)
# backward pass
loss.backward()
current_epoch_loss = loss.item()
# reset the regularizer to free the computational graph
self.model.regularizer.reset()
return current_epoch_loss
@staticmethod
@abstractmethod
def _get_batch_size(batch: Union[MappedTriples, Tuple[MappedTriples, torch.FloatTensor]]) -> int:
"""Get the batch size from a (sub-) batch."""
raise NotImplementedError
@abstractmethod
def _create_instances(self, use_tqdm: Optional[bool] = None) -> Instances:
"""Create the training instances at the beginning of the training loop."""
raise NotImplementedError
@abstractmethod
def _process_batch(
self,
batch: Any,
start: int,
stop: int,
label_smoothing: float = 0.0,
slice_size: Optional[int] = None,
) -> torch.FloatTensor:
"""Process a single batch and returns the loss."""
raise NotImplementedError
def batch_size_search(
self,
batch_size: Optional[int] = None,
) -> Tuple[int, bool]:
"""Find the maximum batch size for training with the current setting.
This method checks how big the batch size can be for the current model with the given training data and the
hardware at hand. If possible, the method will output the determined batch size and a boolean value indicating
that this batch size was successfully evaluated. Otherwise, the output will be batch size 1 and the boolean
value will be False.
:param batch_size:
The batch size to start the search with. If None, set batch_size=num_triples (i.e. full batch training).
:return:
Tuple containing the maximum possible batch size as well as an indicator if the evaluation with that size
was successful.
"""
if batch_size is None:
batch_size = 8192
# Set upper bound
batch_size = min(batch_size, self.triples_factory.num_triples)
reached_max = False
evaluated_once = False
logger.info('Starting batch_size search for training now...')
while True:
logger.debug(f'Trying batch_size={batch_size}.')
try:
self._free_graph_and_cache()
self._train(num_epochs=1, batch_size=batch_size, sub_batch_size=None, only_size_probing=True)
except RuntimeError as runtime_error:
self._free_graph_and_cache()
if not is_cudnn_error(runtime_error) and not is_cuda_oom_error(runtime_error):
raise runtime_error
if batch_size == 1:
logger.debug(f"batch_size={batch_size} does not fit into your memory with these parameters.")
break
reached_max = True
batch_size //= 2
if evaluated_once:
logger.info(f'Concluded batch_size search with batch_size={batch_size}.')
break
logger.debug(f'batch_size={batch_size} was too big, trying less now.')
else:
self._free_graph_and_cache()
if not reached_max and batch_size <= self.triples_factory.num_triples:
batch_size *= 2
else:
logger.info(f'Concluded batch_size search with batch_size={batch_size}.')
evaluated_once = True
break
return batch_size, evaluated_once
def sub_batch_and_slice(self, batch_size: int) -> Tuple[int, int]:
"""Check if sub-batching and/or slicing is necessary to train the model on the hardware at hand."""
sub_batch_size, finished_search, supports_sub_batching = self._sub_batch_size_search(batch_size=batch_size)
# If the sub_batch_size did not finish search with a possibility that fits the hardware, we have to try slicing
if not finished_search:
slice_size = self._slice_size_search(
batch_size=batch_size,
sub_batch_size=sub_batch_size,
supports_sub_batching=supports_sub_batching,
)
else:
slice_size = None
return sub_batch_size, slice_size
@abstractmethod
def _slice_size_search(self, batch_size: int, sub_batch_size: int, supports_sub_batching: bool) -> int:
"""Find the maximum slice size for training with the current setting.
This method finds the biggest slice size to train the model with the given training data and the desired batch
and sub_batch size on the hardware at hand. If even the slice size 1 is too high, it will raise an error.
Otherwise it will return the determined slice size.
:param batch_size:
The batch size to use.
:param sub_batch_size:
The sub-batch size to use.
:param supports_sub_batching:
Indicator if the model supports sub-batching. This is used to create appropriate error messages, if needed.
:return:
The slice_size that allows training the model with the given parameters on this hardware.
:raises MemoryError:
If it is not possible to train the model on the hardware at hand with the given parameters.
"""
raise NotImplementedError
def _sub_batch_size_search(self, batch_size: int) -> Tuple[int, bool, bool]:
"""Find the allowable sub batch size for training with the current setting.
This method checks if it is possible to train the model with the given training data and the desired batch size
on the hardware at hand. If possible, the sub-batch size equals the batch size. Otherwise, the maximum
permissible sub-batch size is determined.
:param batch_size:
The initial batch size to start with.
:return:
Tuple containing the sub-batch size to use and indicating if the search was finished, i.e. successfully
without hardware errors, as well as if sub-batching is possible
"""
sub_batch_size = batch_size
finished_search = False
supports_sub_batching = True
try:
# The cache of the previous run has to be freed to allow accurate memory availability estimates
self._free_graph_and_cache()
logger.debug(f'Trying batch_size {batch_size} for training now.')
self._train(num_epochs=1, batch_size=batch_size, sub_batch_size=sub_batch_size, only_size_probing=True)
except RuntimeError as runtime_error:
self._free_graph_and_cache()
if not is_cudnn_error(runtime_error) and not is_cuda_oom_error(runtime_error):
raise runtime_error
logger.debug(f'The batch_size {batch_size} was too big, sub_batching is required.')
sub_batch_size //= 2
else:
finished_search = True
logger.debug('No sub-batching required.')
if not finished_search:
logger.info('Starting sub_batch_size search for training now...')
if not self.model.supports_subbatching:
logger.info('This model does not support sub-batching.')
supports_sub_batching = False
sub_batch_size = batch_size
else:
while True:
logger.debug(f'Trying sub_batch_size {sub_batch_size} now.')
try:
self._free_graph_and_cache()
self._train(
num_epochs=1,
batch_size=batch_size,
sub_batch_size=sub_batch_size,
only_size_probing=True
)
except RuntimeError as runtime_error:
self._free_graph_and_cache()
if not is_cudnn_error(runtime_error) and not is_cuda_oom_error(runtime_error):
raise runtime_error
if sub_batch_size == 1:
logger.info(
f"Even sub_batch_size={sub_batch_size} does not fit in memory with these parameters")
break
logger.debug(f'The sub_batch_size {sub_batch_size} was too big, trying less now.')
sub_batch_size //= 2
else:
finished_search = True
logger.info(f'Concluded search with sub_batch_size {sub_batch_size}.')
break
self._free_graph_and_cache()
return sub_batch_size, finished_search, supports_sub_batching
def _mr_loss_helper(
self,
positive_scores: torch.FloatTensor,
negative_scores: torch.FloatTensor,
_label_smoothing=None,
) -> torch.FloatTensor:
raise NotImplementedError
def _label_loss_helper(
self,
positive_scores: torch.FloatTensor,
negative_scores: torch.FloatTensor,
label_smoothing: float,
) -> torch.FloatTensor:
raise NotImplementedError
def _self_adversarial_negative_sampling_loss_helper(
self,
positive_scores: torch.FloatTensor,
negative_scores: torch.FloatTensor,
_label_smoothing=None,
) -> torch.FloatTensor:
raise NotImplementedError
def to_embeddingdb(self, session=None, use_tqdm: bool = False):
"""Upload to the embedding database.
:param session: Optional SQLAlchemy session
:param use_tqdm: Use :mod:`tqdm` progress bar?
:rtype: embeddingdb.sql.models.Collection
"""
return self.model.to_embeddingdb(session=session, use_tqdm=use_tqdm)
def _free_graph_and_cache(self):
# The regularizer has to be reset to free the computational graph
self.model.regularizer.reset()
# The cache of the previous run has to be freed to allow accurate memory availability estimates
torch.cuda.empty_cache()