mirror of
https://github.com/wassname/pytorch-lightning.git
synced 2026-09-22 13:30:11 +08:00
Jirka Borovec
d71556e7a1
* upgrade req. * move MkDocs * create Sphinx * init Sphinx * move md from MkDocs to Sphinx * CI: build docs * build Sphinx formatting move docs from MD to docstring in particular package/modules formatting add Sphinx ext. rename root_module to core drop implicit name "_logger" drop duplicate name "overwrite" fix imports use pytorch theme add sample link mapping try fix RTD build use forked template fix some docs warnings fix paths add deprecation warnings fix flake8 fix paths revert refactor revert MLFlowLogger * revert example import * update link * Update lightning_module_template.py
1096 lines
39 KiB
Python
1096 lines
39 KiB
Python
import os
|
|
import warnings
|
|
import collections
|
|
from argparse import Namespace
|
|
|
|
import torch
|
|
import torch.distributed as dist
|
|
|
|
from pytorch_lightning.core.decorators import data_loader
|
|
from pytorch_lightning.core.grads import GradInformation
|
|
from pytorch_lightning.core.hooks import ModelHooks
|
|
from pytorch_lightning.core.memory import ModelSummary
|
|
from pytorch_lightning.core.saving import ModelIO
|
|
from pytorch_lightning.trainer.trainer_io import load_hparams_from_tags_csv
|
|
import logging
|
|
from pytorch_lightning.overrides.data_parallel import LightningDistributedDataParallel
|
|
|
|
|
|
class LightningModule(GradInformation, ModelIO, ModelHooks):
|
|
"""
|
|
A LightningModule has the following properties which you can access at any time
|
|
|
|
**logger**
|
|
A reference to the logger you passed into trainer.
|
|
Passing a logger is optional. If you don't pass one in, Lightning will create one
|
|
for you automatically. This logger saves logs to `/os.getcwd()/lightning_logs`::
|
|
|
|
Trainer(logger=your_logger)
|
|
|
|
|
|
Call it from anywhere in your LightningModule to add metrics, images, etc...
|
|
whatever your logger supports.
|
|
|
|
Here is an example using the TestTubeLogger (which is a wrapper
|
|
on 'PyTorch SummaryWriter <https://pytorch.org/docs/stable/tensorboard.html>`_
|
|
with versioned folder structure).
|
|
|
|
.. code-block:: python
|
|
|
|
# if logger is a tensorboard logger or TestTubeLogger
|
|
self.logger.experiment.add_embedding(...)
|
|
self.logger.experiment.log({'val_loss': 0.9})
|
|
self.logger.experiment.add_scalars(...)
|
|
|
|
|
|
**trainer**
|
|
Last resort access to any state the trainer has.
|
|
Changing certain properties here could affect your training run.
|
|
|
|
.. code-block:: python
|
|
|
|
self.trainer.optimizers
|
|
self.trainer.current_epoch
|
|
...
|
|
|
|
Debugging
|
|
---------
|
|
|
|
The LightningModule also offers these tricks to help debug.
|
|
|
|
**example_input_array**
|
|
|
|
In the LightningModule init, you can set a dummy tensor for this property
|
|
to get a print out of sizes coming into and out of every layer.
|
|
|
|
.. code-block:: python
|
|
|
|
def __init__(self):
|
|
# put the dimensions of the first input to your system
|
|
self.example_input_array = torch.rand(5, 28 * 28)
|
|
|
|
|
|
"""
|
|
|
|
def __init__(self, *args, **kwargs):
|
|
super(LightningModule, self).__init__(*args, **kwargs)
|
|
|
|
#: Current dtype
|
|
self.dtype = torch.FloatTensor
|
|
self.exp_save_path = None
|
|
#: The current epoch
|
|
self.current_epoch = 0
|
|
#: Total training batches seen across all epochs
|
|
self.global_step = 0
|
|
self.loaded_optimizer_states_dict = {}
|
|
self.trainer = None
|
|
self.logger = None
|
|
self.example_input_array = None
|
|
|
|
# track if gpu was requested for checkpointing
|
|
#: True if your model is currently running on GPUs.
|
|
#: Useful to set flags around the LightningModule for different CPU vs GPU behavior.
|
|
self.on_gpu = False
|
|
self.use_dp = False
|
|
self.use_ddp = False
|
|
self.use_ddp2 = False
|
|
self.use_amp = False
|
|
|
|
def forward(self, *args, **kwargs):
|
|
"""
|
|
Expand model in into whatever you need.
|
|
Also need to return the target
|
|
:param x:
|
|
:return:
|
|
"""
|
|
raise NotImplementedError
|
|
|
|
def training_step(self, *args, **kwargs):
|
|
"""return loss, dict with metrics for tqdm
|
|
|
|
:param batch: The output of your dataloader. A tensor, tuple or list
|
|
:param int batch_nb: Integer displaying which batch this is
|
|
:return: dict with loss key and optional log, progress keys
|
|
if implementing training_step, return whatever you need in that step:
|
|
- loss -> tensor scalar [REQUIRED]
|
|
- progress_bar -> Dict for progress bar display. Must have only tensors
|
|
- log -> Dict of metrics to add to logger. Must have only tensors (no images, etc)
|
|
|
|
In this step you'd normally do the forward pass and calculate the loss for a batch.
|
|
You can also do fancier things like multiple forward passes or something specific to your model.
|
|
|
|
Example
|
|
-------
|
|
|
|
.. code-block:: python
|
|
|
|
def training_step(self, batch, batch_nb):
|
|
x, y, z = batch
|
|
|
|
# implement your own
|
|
out = self.forward(x)
|
|
loss = self.loss(out, x)
|
|
|
|
logger_logs = {'training_loss': loss} # optional (MUST ALL BE TENSORS)
|
|
|
|
# if using TestTubeLogger or TensorboardLogger you can nest scalars
|
|
logger_logs = {'losses': logger_logs} # optional (MUST ALL BE TENSORS)
|
|
|
|
output = {
|
|
'loss': loss, # required
|
|
'progress_bar': {'training_loss': loss}, # optional (MUST ALL BE TENSORS)
|
|
'log': logger_logs
|
|
}
|
|
|
|
# return a dict
|
|
return output
|
|
|
|
If you define multiple optimizers, this step will also be called with an additional `optimizer_idx` param.
|
|
|
|
.. code-block:: python
|
|
|
|
# Multiple optimizers (ie: GANs)
|
|
def training_step(self, batch, batch_nb, optimizer_idx):
|
|
if optimizer_idx == 0:
|
|
# do training_step with encoder
|
|
if optimizer_idx == 1:
|
|
# do training_step with decoder
|
|
|
|
|
|
If you add truncated back propagation through time you will also get an additional
|
|
argument with the hidden states of the previous step.
|
|
|
|
.. code-block:: python
|
|
|
|
# Truncated back-propagation through time
|
|
def training_step(self, batch, batch_nb, hiddens):
|
|
# hiddens are the hiddens from the previous truncated backprop step
|
|
|
|
You can also return a -1 instead of a dict to stop the current loop. This is useful
|
|
if you want to break out of the current training epoch early.
|
|
"""
|
|
raise NotImplementedError
|
|
|
|
def training_end(self, *args, **kwargs):
|
|
"""return loss, dict with metrics for tqdm
|
|
|
|
:param outputs: What you return in `training_step`.
|
|
:return dict: dictionary with loss key and optional log, progress keys:
|
|
- loss -> tensor scalar [REQUIRED]
|
|
- progress_bar -> Dict for progress bar display. Must have only tensors
|
|
- log -> Dict of metrics to add to logger. Must have only tensors (no images, etc)
|
|
|
|
In certain cases (dp, ddp2), you might want to use all outputs of every process to do something.
|
|
For instance, if using negative samples, you could run a batch via dp and use ALL the outputs
|
|
for a single softmax across the full batch (ie: the denominator would use the full batch).
|
|
|
|
In this case you should define training_end to perform those calculations.
|
|
|
|
Example
|
|
-------
|
|
|
|
.. code-block:: python
|
|
|
|
# WITHOUT training_end
|
|
# if used in DP or DDP2, this batch is 1/nb_gpus large
|
|
def training_step(self, batch, batch_nb):
|
|
# batch is 1/nb_gpus big
|
|
x, y = batch
|
|
|
|
out = self.forward(x)
|
|
loss = self.softmax(out)
|
|
loss = nce_loss(loss)
|
|
return {'loss': loss}
|
|
|
|
# --------------
|
|
# with training_end to do softmax over the full batch
|
|
def training_step(self, batch, batch_nb):
|
|
# batch is 1/nb_gpus big
|
|
x, y = batch
|
|
|
|
out = self.forward(x)
|
|
return {'out': out}
|
|
|
|
def training_end(self, outputs):
|
|
# this out is now the full size of the batch
|
|
out = outputs['out']
|
|
|
|
# this softmax now uses the full batch size
|
|
loss = self.softmax(out)
|
|
loss = nce_loss(loss)
|
|
return {'loss': loss}
|
|
|
|
If you define multiple optimizers, this step will also be called with an additional `optimizer_idx` param.
|
|
|
|
.. code-block:: python
|
|
|
|
# Multiple optimizers (ie: GANs)
|
|
def training_step(self, batch, batch_nb, optimizer_idx):
|
|
if optimizer_idx == 0:
|
|
# do training_step with encoder
|
|
if optimizer_idx == 1:
|
|
# do training_step with decoder
|
|
|
|
If you add truncated back propagation through time you will also get an additional argument
|
|
with the hidden states of the previous step.
|
|
|
|
.. code-block:: python
|
|
|
|
# Truncated back-propagation through time
|
|
def training_step(self, batch, batch_nb, hiddens):
|
|
# hiddens are the hiddens from the previous truncated backprop step
|
|
|
|
You can also return a -1 instead of a dict to stop the current loop. This is useful if you want to
|
|
break out of the current training epoch early.
|
|
"""
|
|
pass
|
|
|
|
def validation_step(self, *args, **kwargs):
|
|
"""return whatever outputs will need to be aggregated in validation_end
|
|
|
|
:param batch: The output of your dataloader. A tensor, tuple or list
|
|
:param int batch_nb: Integer displaying which batch this is
|
|
:param int dataloader_idx: Integer displaying which dataloader this is (only if multiple val datasets used)
|
|
:return dict: Dict or OrderedDict - passed to the validation_end step
|
|
|
|
.. code-block:: python
|
|
|
|
# if you have one val dataloader:
|
|
def validation_step(self, batch, batch_nb)
|
|
|
|
# if you have multiple val dataloaders:
|
|
def validation_step(self, batch, batch_nb, dataloader_idxdx)
|
|
|
|
If you don't need to validate you don't need to implement this method.
|
|
In this step you'd normally generate examples or calculate anything of interest such as accuracy.
|
|
|
|
When the validation_step is called, the model has been put in eval mode and PyTorch gradients
|
|
have been disabled. At the end of validation, model goes back to training mode and gradients are enabled.
|
|
|
|
The dict you return here will be available in the `validation_end` method.
|
|
|
|
Example
|
|
-------
|
|
|
|
.. code-block:: python
|
|
|
|
# CASE 1: A single validation dataset
|
|
def validation_step(self, batch, batch_nb):
|
|
x, y = batch
|
|
|
|
# implement your own
|
|
out = self.forward(x)
|
|
loss = self.loss(out, y)
|
|
|
|
# log 6 example images
|
|
# or generated text... or whatever
|
|
sample_imgs = x[:6]
|
|
grid = torchvision.utils.make_grid(sample_imgs)
|
|
self.logger.experiment.add_image('example_images', grid, 0)
|
|
|
|
# calculate acc
|
|
labels_hat = torch.argmax(out, dim=1)
|
|
val_acc = torch.sum(y == labels_hat).item() / (len(y) * 1.0)
|
|
|
|
# all optional...
|
|
# return whatever you need for the collation function validation_end
|
|
output = OrderedDict({
|
|
'val_loss': loss_val,
|
|
'val_acc': torch.tensor(val_acc), # everything must be a tensor
|
|
})
|
|
|
|
# return an optional dict
|
|
return output
|
|
|
|
If you pass in multiple validation datasets, validation_step will have an additional argument.
|
|
|
|
.. code-block:: python
|
|
|
|
# CASE 2: multiple validation datasets
|
|
def validation_step(self, batch, batch_nb, dataset_idx):
|
|
# dataset_idx tells you which dataset this is.
|
|
|
|
The `dataset_idx` corresponds to the order of datasets returned in `val_dataloader`.
|
|
"""
|
|
pass
|
|
|
|
def test_step(self, *args, **kwargs):
|
|
"""return whatever outputs will need to be aggregated in test_end
|
|
|
|
:param batch: The output of your dataloader. A tensor, tuple or list
|
|
:param int batch_nb: Integer displaying which batch this is
|
|
:param int dataloader_idx: Integer displaying which dataloader this is (only if multiple test datasets used)
|
|
:return dict: Dict or OrderedDict with metrics to display in progress bar. All keys must be tensors.
|
|
|
|
.. code-block:: python
|
|
|
|
# if you have one test dataloader:
|
|
def test_step(self, batch, batch_nb)
|
|
|
|
# if you have multiple test dataloaders:
|
|
def test_step(self, batch, batch_nb, dataloader_idxdx)
|
|
|
|
|
|
**OPTIONAL**
|
|
If you don't need to test you don't need to implement this method. In this step you'd normally
|
|
generate examples or calculate anything of interest such as accuracy.
|
|
|
|
When the validation_step is called, the model has been put in eval mode and PyTorch gradients
|
|
have been disabled. At the end of validation, model goes back to training mode and gradients are enabled.
|
|
|
|
The dict you return here will be available in the `test_end` method.
|
|
|
|
This function is used when you execute `trainer.test()`.
|
|
|
|
Example
|
|
-------
|
|
|
|
.. code-block:: python
|
|
|
|
# CASE 1: A single test dataset
|
|
def test_step(self, batch, batch_nb):
|
|
x, y = batch
|
|
|
|
# implement your own
|
|
out = self.forward(x)
|
|
loss = self.loss(out, y)
|
|
|
|
# calculate acc
|
|
labels_hat = torch.argmax(out, dim=1)
|
|
test_acc = torch.sum(y == labels_hat).item() / (len(y) * 1.0)
|
|
|
|
# all optional...
|
|
# return whatever you need for the collation function test_end
|
|
output = OrderedDict({
|
|
'test_loss': loss_test,
|
|
'test_acc': torch.tensor(test_acc), # everything must be a tensor
|
|
})
|
|
|
|
# return an optional dict
|
|
return output
|
|
|
|
|
|
If you pass in multiple test datasets, `test_step` will have an additional argument.
|
|
|
|
.. code-block:: python
|
|
|
|
# CASE 2: multiple test datasets
|
|
def test_step(self, batch, batch_nb, dataset_idx):
|
|
# dataset_idx tells you which dataset this is.
|
|
|
|
|
|
The `dataset_idx` corresponds to the order of datasets returned in `test_dataloader`.
|
|
"""
|
|
pass
|
|
|
|
def validation_end(self, outputs):
|
|
"""Outputs has the appended output after each validation step.
|
|
|
|
:param outputs: List of outputs you defined in validation_step, or if there are multiple dataloaders,
|
|
a list containing a list of outputs for each dataloader
|
|
:return dict: Dictionary or OrderedDict with optional:
|
|
progress_bar -> Dict for progress bar display. Must have only tensors
|
|
log -> Dict of metrics to add to logger. Must have only tensors (no images, etc)
|
|
|
|
If you didn't define a validation_step, this won't be called.
|
|
Called at the end of the validation loop with the outputs of validation_step.
|
|
|
|
The outputs here are strictly for the progress bar.
|
|
If you don't need to display anything, don't return anything.
|
|
Any keys present in 'log', 'progress_bar' or the rest of the dictionary
|
|
are available for callbacks to access.
|
|
|
|
Example
|
|
-------
|
|
|
|
With a single dataloader
|
|
|
|
.. code-block:: python
|
|
|
|
def validation_end(self, outputs):
|
|
val_loss_mean = 0
|
|
val_acc_mean = 0
|
|
for output in outputs:
|
|
val_loss_mean += output['val_loss']
|
|
val_acc_mean += output['val_acc']
|
|
|
|
val_loss_mean /= len(outputs)
|
|
val_acc_mean /= len(outputs)
|
|
tqdm_dict = {'val_loss': val_loss_mean.item(), 'val_acc': val_acc_mean.item()}
|
|
|
|
# show val_loss and val_acc in progress bar but only log val_loss
|
|
results = {
|
|
'progress_bar': tqdm_dict,
|
|
'log': {'val_loss': val_loss_mean.item()}
|
|
}
|
|
return results
|
|
|
|
With multiple dataloaders, `outputs` will be a list of lists. The outer list contains
|
|
one entry per dataloader, while the inner list contains the individual outputs of
|
|
each validation step for that dataloader.
|
|
|
|
.. code-block:: python
|
|
|
|
def validation_end(self, outputs):
|
|
val_loss_mean = 0
|
|
val_acc_mean = 0
|
|
i = 0
|
|
for dataloader_outputs in outputs:
|
|
for output in dataloader_outputs:
|
|
val_loss_mean += output['val_loss']
|
|
val_acc_mean += output['val_acc']
|
|
i += 1
|
|
|
|
val_loss_mean /= i
|
|
val_acc_mean /= i
|
|
tqdm_dict = {'val_loss': val_loss_mean.item(), 'val_acc': val_acc_mean.item()}
|
|
|
|
# show val_loss and val_acc in progress bar but only log val_loss
|
|
results = {
|
|
'progress_bar': tqdm_dict,
|
|
'log': {'val_loss': val_loss_mean.item()}
|
|
}
|
|
return results
|
|
|
|
"""
|
|
pass
|
|
|
|
def test_end(self, outputs):
|
|
"""Outputs has the appended output after each test step.
|
|
|
|
:param outputs: List of outputs you defined in test_step, or if there are multiple dataloaders,
|
|
a list containing a list of outputs for each dataloader
|
|
:return dict: Dict of OrderedDict with metrics to display in progress bar
|
|
|
|
If you didn't define a test_step, this won't be called.
|
|
Called at the end of the test step with the output of each test_step.
|
|
The outputs here are strictly for the progress bar.
|
|
If you don't need to display anything, don't return anything.
|
|
|
|
Example
|
|
-------
|
|
|
|
.. code-block:: python
|
|
|
|
def test_end(self, outputs):
|
|
test_loss_mean = 0
|
|
test_acc_mean = 0
|
|
for output in outputs:
|
|
test_loss_mean += output['test_loss']
|
|
test_acc_mean += output['test_acc']
|
|
|
|
test_loss_mean /= len(outputs)
|
|
test_acc_mean /= len(outputs)
|
|
tqdm_dict = {'test_loss': test_loss_mean.item(), 'test_acc': test_acc_mean.item()}
|
|
|
|
# show test_loss and test_acc in progress bar but only log test_loss
|
|
results = {
|
|
'progress_bar': tqdm_dict,
|
|
'log': {'test_loss': val_loss_mean.item()}
|
|
}
|
|
return results
|
|
|
|
With multiple dataloaders, `outputs` will be a list of lists. The outer list contains
|
|
one entry per dataloader, while the inner list contains the individual outputs of
|
|
each validation step for that dataloader.
|
|
|
|
.. code-block:: python
|
|
|
|
def test_end(self, outputs):
|
|
test_loss_mean = 0
|
|
test_acc_mean = 0
|
|
i = 0
|
|
for dataloader_outputs in outputs:
|
|
for output in dataloader_outputs:
|
|
test_loss_mean += output['test_loss']
|
|
test_acc_mean += output['test_acc']
|
|
i += 1
|
|
|
|
test_loss_mean /= i
|
|
test_acc_mean /= i
|
|
tqdm_dict = {'test_loss': test_loss_mean.item(), 'test_acc': test_acc_mean.item()}
|
|
|
|
# show test_loss and test_acc in progress bar but only log test_loss
|
|
results = {
|
|
'progress_bar': tqdm_dict,
|
|
'log': {'test_loss': val_loss_mean.item()}
|
|
}
|
|
return results
|
|
|
|
"""
|
|
pass
|
|
|
|
def configure_ddp(self, model, device_ids):
|
|
"""Override to init DDP in a different way or use your own wrapper.
|
|
|
|
:param model:
|
|
:param device_ids:
|
|
:return: DDP wrapped model
|
|
|
|
Overwrite to define your own DDP implementation init.
|
|
The only requirement is that:
|
|
1. On a validation batch the call goes to model.validation_step.
|
|
2. On a training batch the call goes to model.training_step.
|
|
3. On a testing batch, the call goes to model.test_step
|
|
|
|
.. code-block:: python
|
|
|
|
def configure_ddp(self, model, device_ids):
|
|
# Lightning DDP simply routes to test_step, val_step, etc...
|
|
model = LightningDistributedDataParallel(
|
|
model,
|
|
device_ids=device_ids,
|
|
find_unused_parameters=True
|
|
)
|
|
return model
|
|
|
|
|
|
"""
|
|
model = LightningDistributedDataParallel(
|
|
model,
|
|
device_ids=device_ids,
|
|
find_unused_parameters=True
|
|
)
|
|
return model
|
|
|
|
def init_ddp_connection(self, proc_rank, world_size):
|
|
"""Connect all procs in the world using the env:// init
|
|
Use the first node as the root address
|
|
|
|
Override to init DDP in your own way.
|
|
|
|
.. code-block:: python
|
|
|
|
def init_ddp_connection(self):
|
|
# use slurm job id for the port number
|
|
# guarantees unique ports across jobs from same grid search
|
|
try:
|
|
# use the last 4 numbers in the job id as the id
|
|
default_port = os.environ['SLURM_JOB_ID']
|
|
default_port = default_port[-4:]
|
|
|
|
# all ports should be in the 10k+ range
|
|
default_port = int(default_port) + 15000
|
|
|
|
except Exception as e:
|
|
default_port = 12910
|
|
|
|
# if user gave a port number, use that one instead
|
|
try:
|
|
default_port = os.environ['MASTER_PORT']
|
|
except Exception:
|
|
os.environ['MASTER_PORT'] = str(default_port)
|
|
|
|
# figure out the root node addr
|
|
try:
|
|
root_node = os.environ['SLURM_NODELIST'].split(' ')[0]
|
|
except Exception:
|
|
root_node = '127.0.0.2'
|
|
|
|
root_node = self.trainer.resolve_root_node_address(root_node)
|
|
os.environ['MASTER_ADDR'] = root_node
|
|
dist.init_process_group('nccl', rank=self.proc_rank, world_size=self.world_size)
|
|
|
|
"""
|
|
# use slurm job id for the port number
|
|
# guarantees unique ports across jobs from same grid search
|
|
try:
|
|
# use the last 4 numbers in the job id as the id
|
|
default_port = os.environ['SLURM_JOB_ID']
|
|
default_port = default_port[-4:]
|
|
|
|
# all ports should be in the 10k+ range
|
|
default_port = int(default_port) + 15000
|
|
|
|
except Exception as e:
|
|
default_port = 12910
|
|
|
|
# if user gave a port number, use that one instead
|
|
try:
|
|
default_port = os.environ['MASTER_PORT']
|
|
except Exception:
|
|
os.environ['MASTER_PORT'] = str(default_port)
|
|
|
|
# figure out the root node addr
|
|
try:
|
|
root_node = os.environ['SLURM_NODELIST'].split(' ')[0]
|
|
except Exception:
|
|
root_node = '127.0.0.2'
|
|
|
|
root_node = self.trainer.resolve_root_node_address(root_node)
|
|
os.environ['MASTER_ADDR'] = root_node
|
|
dist.init_process_group('nccl', rank=proc_rank, world_size=world_size)
|
|
|
|
def configure_apex(self, amp, model, optimizers, amp_level):
|
|
"""
|
|
Override to init AMP your own way
|
|
Must return a model and list of optimizers
|
|
:param amp:
|
|
:param model:
|
|
:param optimizers:
|
|
:param amp_level:
|
|
:return: Apex wrapped model and optimizers
|
|
|
|
Overwrite to define your own Apex implementation init.
|
|
|
|
.. code-block:: python
|
|
|
|
def configure_apex(self, amp, model, optimizers, amp_level):
|
|
model, optimizers = amp.initialize(
|
|
model, optimizers, opt_level=amp_level,
|
|
)
|
|
|
|
return model, optimizers
|
|
"""
|
|
model, optimizers = amp.initialize(
|
|
model, optimizers, opt_level=amp_level,
|
|
)
|
|
|
|
return model, optimizers
|
|
|
|
def configure_optimizers(self):
|
|
"""Return a list of optimizers and a list of schedulers (could be empty)
|
|
|
|
:return: any of these 3 options:
|
|
- Single optimizer
|
|
- List or Tuple - List of optimizers
|
|
- Two lists - The first list has multiple optimizers, the second a list of learning-rate schedulers
|
|
|
|
Set up as many optimizers and (optionally) learning rate schedulers as you need.
|
|
Normally you'd need one. But in the case of GANs or something more esoteric you might have multiple.
|
|
Lightning will call .backward() and .step() on each one in every epoch.
|
|
If you use 16 bit precision it will also handle that.
|
|
|
|
.. note:: If you use multiple optimizers, training_step will have an additional `optimizer_idx` parameter.
|
|
|
|
.. note:: If you use LBFGS lightning handles the closure function automatically for you.
|
|
|
|
Example
|
|
-------
|
|
|
|
.. code-block:: python
|
|
|
|
# most cases
|
|
def configure_optimizers(self):
|
|
opt = Adam(self.parameters(), lr=0.01)
|
|
return opt
|
|
|
|
# multiple optimizer case (eg: GAN)
|
|
def configure_optimizers(self):
|
|
generator_opt = Adam(self.model_gen.parameters(), lr=0.01)
|
|
disriminator_opt = Adam(self.model_disc.parameters(), lr=0.02)
|
|
return generator_opt, disriminator_opt
|
|
|
|
# example with learning_rate schedulers
|
|
def configure_optimizers(self):
|
|
generator_opt = Adam(self.model_gen.parameters(), lr=0.01)
|
|
disriminator_opt = Adam(self.model_disc.parameters(), lr=0.02)
|
|
discriminator_sched = CosineAnnealing(discriminator_opt, T_max=10)
|
|
return [generator_opt, disriminator_opt], [discriminator_sched]
|
|
|
|
If you need to control how often those optimizers step or override the default .step() schedule,
|
|
override the `optimizer_step` hook.
|
|
|
|
"""
|
|
raise NotImplementedError
|
|
|
|
def optimizer_step(self, epoch_nb, batch_nb, optimizer, optimizer_i, second_order_closure=None):
|
|
"""Do something instead of the standard optimizer behavior
|
|
|
|
:param int epoch_nb:
|
|
:param int batch_nb:
|
|
:param optimizer:
|
|
:param optimizer_i:
|
|
:param second_order_closure: closure for second order methods
|
|
:return:
|
|
|
|
Calls `.step()` and `.zero_grad` for each optimizer.
|
|
You can override this method to adjust how you do the optimizer step for each optimizer
|
|
|
|
Called once per optimizer
|
|
|
|
.. code-block:: python
|
|
|
|
# DEFAULT
|
|
def optimizer_step(self, current_epoch, batch_nb, optimizer, optimizer_i, second_order_closure=None):
|
|
optimizer.step()
|
|
optimizer.zero_grad()
|
|
|
|
# Alternating schedule for optimizer steps (ie: GANs)
|
|
def optimizer_step(self, current_epoch, batch_nb, optimizer, optimizer_i, second_order_closure=None):
|
|
# update generator opt every 2 steps
|
|
if optimizer_i == 0:
|
|
if batch_nb % 2 == 0 :
|
|
optimizer.step()
|
|
optimizer.zero_grad()
|
|
|
|
# update discriminator opt every 4 steps
|
|
if optimizer_i == 1:
|
|
if batch_nb % 4 == 0 :
|
|
optimizer.step()
|
|
optimizer.zero_grad()
|
|
|
|
# ...
|
|
# add as many optimizers as you want
|
|
|
|
|
|
This step allows you to do a lot of non-standard training tricks such as learning-rate warm-up:
|
|
|
|
.. code-block:: python
|
|
|
|
# learning rate warm-up
|
|
def optimizer_step(self, current_epoch, batch_nb, optimizer, optimizer_i, second_order_closure=None):
|
|
# warm up lr
|
|
if self.trainer.global_step < 500:
|
|
lr_scale = min(1., float(self.trainer.global_step + 1) / 500.)
|
|
for pg in optimizer.param_groups:
|
|
pg['lr'] = lr_scale * self.hparams.learning_rate
|
|
|
|
# update params
|
|
optimizer.step()
|
|
optimizer.zero_grad()
|
|
|
|
"""
|
|
if isinstance(optimizer, torch.optim.LBFGS):
|
|
optimizer.step(second_order_closure)
|
|
else:
|
|
optimizer.step()
|
|
|
|
# clear gradients
|
|
optimizer.zero_grad()
|
|
|
|
def tbptt_split_batch(self, batch, split_size):
|
|
"""
|
|
Return list of batch splits. Each split will be passed to forward_step to enable truncated
|
|
back propagation through time. The default implementation splits root level Tensors and
|
|
Sequences at dim=1 (i.e. time dim). It assumes that each time dim is the same length.
|
|
|
|
:param batch:
|
|
:param split_size:
|
|
:return:
|
|
|
|
Called in the training loop after on_batch_start if `truncated_bptt_steps > 0`.
|
|
Each returned batch split is passed separately to training_step(...).
|
|
|
|
.. code-block:: python
|
|
|
|
def tbptt_split_batch(self, batch, split_size):
|
|
splits = []
|
|
for t in range(0, time_dims[0], split_size):
|
|
batch_split = []
|
|
for i, x in enumerate(batch):
|
|
if isinstance(x, torch.Tensor):
|
|
split_x = x[:, t:t + split_size]
|
|
elif isinstance(x, collections.Sequence):
|
|
split_x = [None] * len(x)
|
|
for batch_idx in range(len(x)):
|
|
split_x[batch_idx] = x[batch_idx][t:t + split_size]
|
|
|
|
batch_split.append(split_x)
|
|
|
|
splits.append(batch_split)
|
|
|
|
return splits
|
|
"""
|
|
time_dims = [len(x[0]) for x in batch if isinstance(
|
|
x, torch.Tensor) or isinstance(x, collections.Sequence)]
|
|
assert len(time_dims) >= 1, "Unable to determine batch time dimension"
|
|
assert all(x == time_dims[0] for x in time_dims), "Batch time dimension length is ambiguous"
|
|
|
|
splits = []
|
|
for t in range(0, time_dims[0], split_size):
|
|
batch_split = []
|
|
for i, x in enumerate(batch):
|
|
if isinstance(x, torch.Tensor):
|
|
split_x = x[:, t:t + split_size]
|
|
elif isinstance(x, collections.Sequence):
|
|
split_x = [None] * len(x)
|
|
for batch_idx in range(len(x)):
|
|
split_x[batch_idx] = x[batch_idx][t:t + split_size]
|
|
|
|
batch_split.append(split_x)
|
|
|
|
splits.append(batch_split)
|
|
|
|
return splits
|
|
|
|
@data_loader
|
|
def tng_dataloader(self):
|
|
"""
|
|
Implement a PyTorch DataLoader
|
|
* Deprecated in v0.5.0. use train_dataloader instead. *
|
|
:return:
|
|
"""
|
|
raise NotImplementedError
|
|
|
|
@data_loader
|
|
def train_dataloader(self):
|
|
"""Implement a PyTorch DataLoader
|
|
|
|
:return: PyTorch DataLoader
|
|
|
|
Called by lightning during training loop. Make sure to use the @pl.data_loader decorator,
|
|
this ensures not calling this function until the data are needed.
|
|
If you want to change the data during every epoch DON'T use the data_loader decorator.
|
|
|
|
Example
|
|
-------
|
|
|
|
.. code-block:: python
|
|
|
|
@pl.data_loader
|
|
def train_dataloader(self):
|
|
transform = transforms.Compose([transforms.ToTensor(), transforms.Normalize((0.5,), (1.0,))])
|
|
dataset = MNIST(root='/path/to/mnist/', train=True, transform=transform, download=True)
|
|
loader = torch.utils.data.DataLoader(
|
|
dataset=dataset,
|
|
batch_size=self.hparams.batch_size,
|
|
shuffle=True
|
|
)
|
|
return loader
|
|
|
|
|
|
"""
|
|
#
|
|
try:
|
|
output = self.tng_dataloader()
|
|
warnings.warn("`tng_dataloader` has been renamed to `train_dataloader` since v0.5.0"
|
|
" and will be removed in v0.8.0", DeprecationWarning)
|
|
return output
|
|
except NotImplementedError:
|
|
raise NotImplementedError
|
|
|
|
@data_loader
|
|
def test_dataloader(self):
|
|
"""Implement a PyTorch DataLoader.
|
|
|
|
:return: PyTorch DataLoader
|
|
|
|
If you don't need a test dataset and a test_step, you don't need to implement this method.
|
|
|
|
Called by lightning during test loop. Make sure to use the @pl.data_loader decorator,
|
|
this ensures not calling this function until the data are needed.
|
|
If you want to change the data during every epoch DON'T use the data_loader decorator.
|
|
|
|
Example
|
|
-------
|
|
|
|
.. code-block:: python
|
|
|
|
@pl.data_loader
|
|
def test_dataloader(self):
|
|
transform = transforms.Compose([transforms.ToTensor(), transforms.Normalize((0.5,), (1.0,))])
|
|
dataset = MNIST(root='/path/to/mnist/', train=False, transform=transform, download=True)
|
|
loader = torch.utils.data.DataLoader(
|
|
dataset=dataset,
|
|
batch_size=self.hparams.batch_size,
|
|
shuffle=True
|
|
)
|
|
|
|
return loader
|
|
|
|
"""
|
|
return None
|
|
|
|
@data_loader
|
|
def val_dataloader(self):
|
|
"""Implement a PyTorch DataLoader.
|
|
|
|
:return: PyTorch DataLoader or list of PyTorch Dataloaders.
|
|
|
|
If you don't need a validation dataset and a validation_step, you don't need to implement this method.
|
|
|
|
Called by lightning during validation loop. Make sure to use the @pl.data_loader decorator,
|
|
this ensures not calling this function until the data are needed.
|
|
If you want to change the data during every epoch DON'T use the data_loader decorator.
|
|
|
|
Example
|
|
-------
|
|
|
|
.. code-block:: python
|
|
|
|
@pl.data_loader
|
|
def val_dataloader(self):
|
|
transform = transforms.Compose([transforms.ToTensor(), transforms.Normalize((0.5,), (1.0,))])
|
|
dataset = MNIST(root='/path/to/mnist/', train=False, transform=transform, download=True)
|
|
loader = torch.utils.data.DataLoader(
|
|
dataset=dataset,
|
|
batch_size=self.hparams.batch_size,
|
|
shuffle=True
|
|
)
|
|
|
|
return loader
|
|
|
|
# can also return multiple dataloaders
|
|
@pl.data_loader
|
|
def val_dataloader(self):
|
|
return [loader_a, loader_b, ..., loader_n]
|
|
|
|
In the case where you return multiple `val_dataloaders`, the `validation_step`
|
|
will have an arguement `dataset_idx` which matches the order here.
|
|
"""
|
|
return None
|
|
|
|
@classmethod
|
|
def load_from_metrics(cls, weights_path, tags_csv):
|
|
"""Primary way of loading model from csv weights path.
|
|
|
|
:param str weights_path: Path to a PyTorch checkpoint
|
|
:param str tags_csv: Path to meta_tags.csv file generated by the test-tube Experiment
|
|
:param dict map_location: A dictionary mapping saved weight GPU devices to new GPU devices
|
|
for mapping storage {'cuda:1':'cuda:0'}
|
|
:return: The pretrained LightningModule
|
|
|
|
If you're using test tube, there is an alternate method which uses the meta_tags.csv
|
|
file from test-tube to rebuild the model. The meta_tags.csv file can be found in the
|
|
test-tube experiment save_dir.
|
|
|
|
.. code-block:: python
|
|
|
|
pretrained_model = MyLightningModule.load_from_metrics(
|
|
weights_path='/path/to/pytorch_checkpoint.ckpt',
|
|
tags_csv='/path/to/test_tube/experiment/version/meta_tags.csv',
|
|
on_gpu=True,
|
|
map_location=None
|
|
)
|
|
|
|
# predict
|
|
pretrained_model.eval()
|
|
pretrained_model.freeze()
|
|
y_hat = pretrained_model(x)
|
|
|
|
This is the easiest/fastest way which loads hyperparameters and weights from a checkpoint,
|
|
such as the one saved by the `ModelCheckpoint` callback
|
|
|
|
.. code-block:: python
|
|
|
|
pretrained_model = MyLightningModule.load_from_checkpoint(
|
|
checkpoint_path='/path/to/pytorch_checkpoint.ckpt'
|
|
)
|
|
|
|
# predict
|
|
pretrained_model.eval()
|
|
pretrained_model.freeze()
|
|
y_hat = pretrained_model(x)
|
|
|
|
"""
|
|
hparams = load_hparams_from_tags_csv(tags_csv)
|
|
hparams.__setattr__('on_gpu', False)
|
|
|
|
# load on CPU only to avoid OOM issues
|
|
# then its up to user to put back on GPUs
|
|
checkpoint = torch.load(weights_path, map_location=lambda storage, loc: storage)
|
|
|
|
# load the state_dict on the model automatically
|
|
model = cls(hparams)
|
|
model.load_state_dict(checkpoint['state_dict'])
|
|
|
|
# give model a chance to load something
|
|
model.on_load_checkpoint(checkpoint)
|
|
|
|
return model
|
|
|
|
@classmethod
|
|
def load_from_checkpoint(cls, checkpoint_path):
|
|
"""
|
|
Primary way of loading model from a checkpoint
|
|
:param checkpoint_path:
|
|
:param map_location: dic for mapping storage {'cuda:1':'cuda:0'}
|
|
:return:
|
|
"""
|
|
|
|
# load on CPU only to avoid OOM issues
|
|
# then its up to user to put back on GPUs
|
|
checkpoint = torch.load(checkpoint_path, map_location=lambda storage, loc: storage)
|
|
try:
|
|
ckpt_hparams = checkpoint['hparams']
|
|
except KeyError:
|
|
raise IOError(
|
|
"Checkpoint does not contain hyperparameters. Are your model hyperparameters stored"
|
|
"in self.hparams?"
|
|
)
|
|
hparams = Namespace(**ckpt_hparams)
|
|
|
|
# load the state_dict on the model automatically
|
|
model = cls(hparams)
|
|
model.load_state_dict(checkpoint['state_dict'])
|
|
|
|
# give model a chance to load something
|
|
model.on_load_checkpoint(checkpoint)
|
|
|
|
return model
|
|
|
|
def summarize(self, mode):
|
|
model_summary = ModelSummary(self, mode=mode)
|
|
logging.info(model_summary)
|
|
|
|
def freeze(self):
|
|
"""Freeze all params for inference
|
|
|
|
.. code-block:: python
|
|
|
|
model = MyLightningModule(...)
|
|
model.freeze()
|
|
|
|
"""
|
|
for param in self.parameters():
|
|
param.requires_grad = False
|
|
|
|
self.eval()
|
|
|
|
def unfreeze(self):
|
|
"""Unfreeze all params for inference.
|
|
|
|
.. code-block:: python
|
|
|
|
model = MyLightningModule(...)
|
|
model.unfreeze()
|
|
|
|
"""
|
|
for param in self.parameters():
|
|
param.requires_grad = True
|
|
|
|
self.train()
|
|
|
|
def on_load_checkpoint(self, checkpoint):
|
|
"""
|
|
|
|
:param checkpoint:
|
|
|
|
Called by lightning to restore your model. Lighting auto-restores global step, epoch, etc...
|
|
It also restores the model state_dict.
|
|
If you saved something with **on_save_checkpoint** this is your chance to restore this.
|
|
|
|
Example
|
|
-------
|
|
|
|
.. code-block:: python
|
|
|
|
def on_load_checkpoint(self, checkpoint):
|
|
# 99% of the time you don't need to implement this method
|
|
self.something_cool_i_want_to_save = checkpoint['something_cool_i_want_to_save']
|
|
|
|
"""
|
|
pass
|
|
|
|
def on_save_checkpoint(self, checkpoint):
|
|
"""
|
|
|
|
:param checkpoint:
|
|
|
|
Called by lightning to checkpoint your model. Lightning saves the training state
|
|
(current epoch, global_step, etc) and also saves the model state_dict.
|
|
If you want to save anything else, use this method to add your own key-value pair.
|
|
|
|
Example
|
|
-------
|
|
|
|
.. code-block:: python
|
|
|
|
def on_save_checkpoint(self, checkpoint):
|
|
# 99% of use cases you don't need to implement this method
|
|
checkpoint['something_cool_i_want_to_save'] = my_cool_pickable_object
|
|
|
|
"""
|
|
pass
|