lightning/pytorch_lightning/trainer/train_loop_mixin.py

461 lines
16 KiB
Python

"""
The lightning training loop handles everything except the actual computations of your model.
To decide what will happen in your training loop, define the `training_step` function.
Below are all the things lightning automates for you in the training loop.
Accumulated gradients
---------------------
Accumulated gradients runs K small batches of size N before doing a backwards pass.
The effect is a large effective batch size of size KxN.
.. code-block:: python
# DEFAULT (ie: no accumulated grads)
trainer = Trainer(accumulate_grad_batches=1)
Force training for min or max epochs
------------------------------------
It can be useful to force training for a minimum number of epochs or limit to a max number
.. code-block:: python
# DEFAULT
trainer = Trainer(min_nb_epochs=1, max_nb_epochs=1000)
Early stopping
--------------
The trainer already sets up default early stopping for you.
To modify this behavior, pass in your own EarlyStopping callback.
.. code-block:: python
from pytorch_lightning.callbacks import EarlyStopping
# DEFAULTS used by Trainer
early_stop_callback = EarlyStopping(
monitor='val_loss',
min_delta=0.00,
patience=3,
verbose=False,
mode='min'
)
# without passing anything in, uses the default callback above
trainer = Trainer()
# pass in your own to override the default callback
trainer = Trainer(early_stop_callback=early_stop_callback)
# pass in None to disable it
trainer = Trainer(early_stop_callback=None)
Force disable early stop
------------------------
To disable early stopping pass None to the early_stop_callback
.. code-block:: python
# DEFAULT
trainer = Trainer(early_stop_callback=None)
Gradient Clipping
-----------------
Gradient clipping may be enabled to avoid exploding gradients.
Specifically, this will `clip the gradient norm computed over all model parameters
`together <https://pytorch.org/docs/stable/nn.html#torch.nn.utils.clip_grad_norm_>`_.
.. code-block:: python
# DEFAULT (ie: don't clip)
trainer = Trainer(gradient_clip_val=0)
# clip gradients with norm above 0.5
trainer = Trainer(gradient_clip_val=0.5)
Inspect gradient norms
----------------------
Looking at grad norms can help you figure out where training might be going wrong.
.. code-block:: python
# DEFAULT (-1 doesn't track norms)
trainer = Trainer(track_grad_norm=-1)
# track the LP norm (P=2 here)
trainer = Trainer(track_grad_norm=2)
Set how much of the training set to check
-----------------------------------------
If you don't want to check 100% of the training set (for debugging or if it's huge), set this flag.
train_percent_check will be overwritten by overfit_pct if `overfit_pct > 0`
.. code-block:: python
# DEFAULT
trainer = Trainer(train_percent_check=1.0)
# check 10% only
trainer = Trainer(train_percent_check=0.1)
Packed sequences as inputs
--------------------------
When using PackedSequence, do 2 things:
1. return either a padded tensor in dataset or a list of variable length tensors
in the dataloader collate_fn (example above shows the list implementation).
2. Pack the sequence in forward or training and validation steps depending on use case.
.. code-block:: python
# For use in dataloader
def collate_fn(batch):
x = [item[0] for item in batch]
y = [item[1] for item in batch]
return x, y
# In module
def training_step(self, batch, batch_nb):
x = rnn.pack_sequence(batch[0], enforce_sorted=False)
y = rnn.pack_sequence(batch[1], enforce_sorted=False)
Truncated Backpropagation Through Time
--------------------------------------
There are times when multiple backwards passes are needed for each batch.
For example, it may save memory to use Truncated Backpropagation Through Time when training RNNs.
When this flag is enabled each batch is split into sequences of size truncated_bptt_steps
and passed to training_step(...) separately. A default splitting function is provided,
however, you can override it for more flexibility. See `tbptt_split_batch`.
.. code-block:: python
# DEFAULT (single backwards pass per batch)
trainer = Trainer(truncated_bptt_steps=None)
# (split batch into sequences of size 2)
trainer = Trainer(truncated_bptt_steps=2)
"""
import numpy as np
import tqdm
try:
from apex import amp
APEX_AVAILABLE = True
except ImportError:
APEX_AVAILABLE = False
class TrainerTrainLoopMixin(object):
def train(self):
# run all epochs
for epoch_nb in range(self.current_epoch, self.max_nb_epochs):
# set seed for distributed sampler (enables shuffling for each epoch)
if self.use_ddp and hasattr(self.get_train_dataloader().sampler, 'set_epoch'):
self.get_train_dataloader().sampler.set_epoch(epoch_nb)
# get model
model = self.get_model()
# update training progress in trainer and model
model.current_epoch = epoch_nb
self.current_epoch = epoch_nb
# val can be checked multiple times in epoch
is_val_epoch = (self.current_epoch + 1) % self.check_val_every_n_epoch == 0
val_checks_per_epoch = self.nb_training_batches // self.val_check_batch
val_checks_per_epoch = val_checks_per_epoch if is_val_epoch else 0
# total batches includes multiple val checks
self.total_batches = (self.nb_training_batches +
self.nb_val_batches * val_checks_per_epoch)
self.batch_loss_value = 0 # accumulated grads
if self.fast_dev_run:
# limit the number of batches to 2 (1 train and 1 val) in fast_dev_run
nb_iterations = 2
elif self.is_iterable_train_dataloader:
# for iterable train loader, the progress bar never ends
nb_iterations = None
else:
nb_iterations = self.total_batches
# reset progress bar
# .reset() doesn't work on disabled progress bar so we should check
if not self.main_progress_bar.disable:
self.main_progress_bar.reset(nb_iterations)
desc = f'Epoch {epoch_nb + 1}' if not self.is_iterable_train_dataloader else ''
self.main_progress_bar.set_description(desc)
# changing gradient according accumulation_scheduler
self.accumulation_scheduler.on_epoch_begin(epoch_nb, self)
# -----------------
# RUN TNG EPOCH
# -----------------
self.run_training_epoch()
# update LR schedulers
if self.lr_schedulers is not None:
for lr_scheduler in self.lr_schedulers:
lr_scheduler.step(self.current_epoch)
# early stopping
met_min_epochs = epoch_nb > self.min_nb_epochs
if self.enable_early_stop and (met_min_epochs or self.fast_dev_run):
should_stop = self.early_stop_callback.on_epoch_end(epoch=epoch_nb,
logs=self.callback_metrics)
# stop training
stop = should_stop and met_min_epochs
if stop:
self.main_progress_bar.close()
return
self.main_progress_bar.close()
if self.logger is not None:
self.logger.finalize("success")
def run_training_epoch(self):
# before epoch hook
if self.is_function_implemented('on_epoch_start'):
model = self.get_model()
model.on_epoch_start()
# run epoch
for batch_nb, batch in enumerate(self.get_train_dataloader()):
self.batch_nb = batch_nb
model = self.get_model()
model.global_step = self.global_step
# ---------------
# RUN TRAIN STEP
# ---------------
output = self.run_training_batch(batch, batch_nb)
batch_result, grad_norm_dic, batch_step_metrics = output
# when returning -1 from train_step, we end epoch early
early_stop_epoch = batch_result == -1
# ---------------
# RUN VAL STEP
# ---------------
is_val_check_batch = (batch_nb + 1) % self.val_check_batch == 0
can_check_epoch = (self.current_epoch + 1) % self.check_val_every_n_epoch == 0
should_check_val = ((is_val_check_batch or early_stop_epoch) and can_check_epoch)
# fast_dev_run always forces val checking after train batch
if self.fast_dev_run or should_check_val:
self.run_evaluation(test=self.testing)
# when logs should be saved
should_save_log = (batch_nb + 1) % self.log_save_interval == 0 or early_stop_epoch
if should_save_log or self.fast_dev_run:
if self.proc_rank == 0 and self.logger is not None:
self.logger.save()
# when metrics should be logged
should_log_metrics = batch_nb % self.row_log_interval == 0 or early_stop_epoch
if should_log_metrics or self.fast_dev_run:
# logs user requested information to logger
self.log_metrics(batch_step_metrics, grad_norm_dic)
self.global_step += 1
self.total_batch_nb += 1
# end epoch early
# stop when the flag is changed or we've gone past the amount
# requested in the batches
if early_stop_epoch or self.fast_dev_run:
break
# stop epoch if we limited nb batches
met_batch_limit = batch_nb >= self.nb_training_batches
if met_batch_limit:
break
# epoch end hook
if self.is_function_implemented('on_epoch_end'):
model = self.get_model()
model.on_epoch_end()
def run_training_batch(self, batch, batch_nb):
# track grad norms
grad_norm_dic = {}
# track all metrics for callbacks
all_callback_metrics = []
# track metrics to log
all_log_metrics = []
if batch is None:
return 0, grad_norm_dic, {}
# hook
if self.is_function_implemented('on_batch_start'):
model_ref = self.get_model()
response = model_ref.on_batch_start(batch)
if response == -1:
return -1, grad_norm_dic, {}
splits = [batch]
if self.truncated_bptt_steps is not None:
model_ref = self.get_model()
splits = model_ref.tbptt_split_batch(batch, self.truncated_bptt_steps)
self.hiddens = None
for split_nb, split_batch in enumerate(splits):
self.split_nb = split_nb
# call training_step once per optimizer
for opt_idx, optimizer in enumerate(self.optimizers):
# wrap the forward step in a closure so second order methods work
def optimizer_closure():
# forward pass
output = self.training_forward(
split_batch, batch_nb, opt_idx, self.hiddens)
closure_loss = output[0]
progress_bar_metrics = output[1]
log_metrics = output[2]
callback_metrics = output[3]
self.hiddens = output[4]
# accumulate loss
# (if accumulate_grad_batches = 1 no effect)
closure_loss = closure_loss / self.accumulate_grad_batches
# backward pass
model_ref = self.get_model()
model_ref.backward(self.use_amp, closure_loss, optimizer)
# track metrics for callbacks
all_callback_metrics.append(callback_metrics)
# track progress bar metrics
self.add_tqdm_metrics(progress_bar_metrics)
all_log_metrics.append(log_metrics)
# insert after step hook
if self.is_function_implemented('on_after_backward'):
model_ref = self.get_model()
model_ref.on_after_backward()
return closure_loss
# calculate loss
loss = optimizer_closure()
# nan grads
if self.print_nan_grads:
self.print_nan_gradients()
# track total loss for logging (avoid mem leaks)
self.batch_loss_value += loss.item()
# gradient update with accumulated gradients
if (self.batch_nb + 1) % self.accumulate_grad_batches == 0:
# track gradient norms when requested
if batch_nb % self.row_log_interval == 0:
if self.track_grad_norm > 0:
model = self.get_model()
grad_norm_dic = model.grad_norm(
self.track_grad_norm)
# clip gradients
self.clip_gradients()
# calls .step(), .zero_grad()
# override function to modify this behavior
model = self.get_model()
model.optimizer_step(self.current_epoch, batch_nb,
optimizer, opt_idx, optimizer_closure)
# calculate running loss for display
self.running_loss.append(self.batch_loss_value)
self.batch_loss_value = 0
self.avg_loss = np.mean(self.running_loss[-100:])
# activate batch end hook
if self.is_function_implemented('on_batch_end'):
model = self.get_model()
model.on_batch_end()
# update progress bar
self.main_progress_bar.update(1)
self.main_progress_bar.set_postfix(**self.training_tqdm_dict)
# collapse all metrics into one dict
all_log_metrics = {k: v for d in all_log_metrics for k, v in d.items()}
# track all metrics for callbacks
self.callback_metrics.update({k: v for d in all_callback_metrics for k, v in d.items()})
return 0, grad_norm_dic, all_log_metrics
def training_forward(self, batch, batch_nb, opt_idx, hiddens):
"""
Handle forward for each training case (distributed, single gpu, etc...)
:param batch:
:param batch_nb:
:return:
"""
# ---------------
# FORWARD
# ---------------
# enable not needing to add opt_idx to training_step
args = [batch, batch_nb]
if len(self.optimizers) > 1:
args.append(opt_idx)
# pass hiddens if using tbptt
if self.truncated_bptt_steps is not None:
args.append(hiddens)
# distributed forward
if self.use_ddp or self.use_ddp2 or self.use_dp:
output = self.model(*args)
# single GPU forward
elif self.single_gpu:
gpu_id = 0
if type(self.data_parallel_device_ids) is list:
gpu_id = self.data_parallel_device_ids[0]
batch = self.transfer_batch_to_gpu(batch.copy(), gpu_id)
args[0] = batch
output = self.model.training_step(*args)
# CPU forward
else:
output = self.model.training_step(*args)
# allow any mode to define training_end
if self.is_overriden('training_end'):
model_ref = self.get_model()
output = model_ref.training_end(output)
# format and reduce outputs accordingly
output = self.process_output(output, train=True)
return output