"""
Block Recurrent Neural Networks
-------------------------------
"""
import inspect
from abc import ABC, abstractmethod
from typing import List, Optional, Tuple, Type, Union
import torch
import torch.nn as nn
from darts.logging import get_logger, raise_log
from darts.models.forecasting.pl_forecasting_module import (
PLPastCovariatesModule,
io_processor,
)
from darts.models.forecasting.torch_forecasting_model import PastCovariatesTorchModel
logger = get_logger(__name__)
[docs]class CustomBlockRNNModule(PLPastCovariatesModule, ABC):
def __init__(
self,
input_size: int,
hidden_dim: int,
num_layers: int,
target_size: int,
nr_params: int,
num_layers_out_fc: Optional[List] = None,
dropout: float = 0.0,
**kwargs,
):
"""This class allows to create custom block RNN modules that can later be used with Darts'
:class:`BlockRNNModel`. It adds the backbone that is required to be used with Darts'
:class:`TorchForecastingModel` and :class:`BlockRNNModel`.
To create a new module, subclass from :class:`CustomBlockRNNModule` and:
* Define the architecture in the module constructor (`__init__()`)
* Add the `forward()` method and define the logic of your module's forward pass
* Use the custom module class when creating a new :class:`BlockRNNModel` with parameter `model`.
You can use `darts.models.forecasting.block_rnn_model._BlockRNNModule` as an example.
Parameters
----------
input_size
The dimensionality of the input time series.
hidden_dim
The number of features in the hidden state `h` of the RNN module.
num_layers
The number of recurrent layers.
target_size
The dimensionality of the output time series.
nr_params
The number of parameters of the likelihood (or 1 if no likelihood is used).
num_layers_out_fc
A list containing the dimensions of the hidden layers of the fully connected NN.
This network connects the last hidden layer of the PyTorch RNN module to the output.
dropout
The fraction of neurons that are dropped in all-but-last RNN layers.
**kwargs
all parameters required for :class:`darts.models.forecasting.pl_forecasting_module.PLForecastingModule`
base class.
"""
super().__init__(**kwargs)
# Defining parameters
self.input_size = input_size
self.hidden_dim = hidden_dim
self.num_layers = num_layers
self.target_size = target_size
self.nr_params = nr_params
self.num_layers_out_fc = [] if num_layers_out_fc is None else num_layers_out_fc
self.dropout = dropout
self.out_len = self.output_chunk_length
[docs] @io_processor
@abstractmethod
def forward(self, x_in: Tuple) -> torch.Tensor:
"""BlockRNN Module forward.
Parameters
----------
x_in
Tuple of Tensors containing the features of the input sequence. The tuple has elements
(past target, historic future covariates, future covariates, static covariates).
The shape of the past target is `(batch_size, input_length, input_size)`.
Returns
-------
torch.Tensor
The BlockRNN output Tensor with shape `(batch_size, output_chunk_length, target_size, nr_params)`.
It contains the prediction at the last time step of the sequence.
"""
pass
# TODO add batch norm
class _BlockRNNModule(CustomBlockRNNModule):
def __init__(
self,
name: str,
**kwargs,
):
"""PyTorch module implementing a block RNN to be used in `BlockRNNModel`.
PyTorch module implementing a simple block RNN with the specified `name` layer.
This module combines a PyTorch RNN module, together with a fully connected network, which maps the
last hidden layers to output of the desired size `output_chunk_length` and makes it compatible with
`BlockRNNModel`s.
This module uses an RNN to encode the input sequence, and subsequently uses a fully connected
network as the decoder which takes as input the last hidden state of the encoder RNN.
The final output of the decoder is a sequence of length `output_chunk_length`. In this sense,
the `_BlockRNNModule` produces 'blocks' of forecasts at a time (which is different
from `_RNNModule` used by the `RNNModel`).
Parameters
----------
name
The name of the specific PyTorch RNN module ("RNN", "GRU" or "LSTM").
**kwargs
all parameters required for the :class:`darts.models.forecasting.CustomBlockRNNModule` base class.
Inputs
------
x of shape `(batch_size, input_chunk_length, input_size, nr_params)`
Tensor containing the features of the input sequence.
Outputs
-------
y of shape `(batch_size, output_chunk_length, target_size, nr_params)`
Tensor containing the prediction at the last time step of the sequence.
"""
super().__init__(**kwargs)
self.name = name
# Defining the RNN module
self.rnn = getattr(nn, self.name)(
self.input_size,
self.hidden_dim,
self.num_layers,
batch_first=True,
dropout=self.dropout,
)
# The RNN module is followed by a fully connected layer, which maps the last hidden layer
# to the output of desired length
last = self.hidden_dim
feats = []
for feature in self.num_layers_out_fc + [
self.out_len * self.target_size * self.nr_params
]:
feats.append(nn.Linear(last, feature))
last = feature
self.fc = nn.Sequential(*feats)
@io_processor
def forward(self, x_in: Tuple):
x, _ = x_in
# data is of size (batch_size, input_chunk_length, input_size)
batch_size = x.size(0)
out, hidden = self.rnn(x)
""" Here, we apply the FC network only on the last output point (at the last time step)
"""
if self.name == "LSTM":
hidden = hidden[0]
predictions = hidden[-1, :, :]
predictions = self.fc(predictions)
predictions = predictions.view(
batch_size, self.out_len, self.target_size, self.nr_params
)
# predictions is of size (batch_size, output_chunk_length, 1)
return predictions
[docs]class BlockRNNModel(PastCovariatesTorchModel):
def __init__(
self,
input_chunk_length: int,
output_chunk_length: int,
output_chunk_shift: int = 0,
model: Union[str, Type[CustomBlockRNNModule]] = "RNN",
hidden_dim: int = 25,
n_rnn_layers: int = 1,
hidden_fc_sizes: Optional[List] = None,
dropout: float = 0.0,
**kwargs,
):
"""Block Recurrent Neural Network Model (RNNs).
This is a neural network model that uses an RNN encoder to encode fixed-length input chunks, and
a fully connected network to produce fixed-length outputs.
This model supports past covariates (known for `input_chunk_length` points before prediction time).
This class provides three variants of RNNs:
* Vanilla RNN
* LSTM
* GRU
Parameters
----------
input_chunk_length
Number of time steps in the past to take as a model input (per chunk). Applies to the target
series, and past and/or future covariates (if the model supports it).
output_chunk_length
Number of time steps predicted at once (per chunk) by the internal model. Also, the number of future values
from future covariates to use as a model input (if the model supports future covariates). It is not the same
as forecast horizon `n` used in `predict()`, which is the desired number of prediction points generated
using either a one-shot- or autoregressive forecast. Setting `n <= output_chunk_length` prevents
auto-regression. This is useful when the covariates don't extend far enough into the future, or to prohibit
the model from using future values of past and / or future covariates for prediction (depending on the
model's covariate support).
output_chunk_shift
Optionally, the number of steps to shift the start of the output chunk into the future (relative to the
input chunk end). This will create a gap between the input and output. If the model supports
`future_covariates`, the future values are extracted from the shifted output chunk. Predictions will start
`output_chunk_shift` steps after the end of the target `series`. If `output_chunk_shift` is set, the model
cannot generate autoregressive predictions (`n > output_chunk_length`).
model
Either a string specifying the RNN module type ("RNN", "LSTM" or "GRU"), or a subclass of
:class:`CustomBlockRNNModule` (the class itself, not an object of the class) with a custom logic.
hidden_dim
Size for feature maps for each hidden RNN layer (:math:`h_n`).
In Darts version <= 0.21, hidden_dim was referred as hidden_size.
n_rnn_layers
Number of layers in the RNN module.
hidden_fc_sizes
Sizes of hidden layers connecting the last hidden layer of the RNN module to the output, if any.
dropout
Fraction of neurons afected by Dropout.
**kwargs
Optional arguments to initialize the pytorch_lightning.Module, pytorch_lightning.Trainer, and
Darts' :class:`TorchForecastingModel`.
loss_fn
PyTorch loss function used for training.
This parameter will be ignored for probabilistic models if the ``likelihood`` parameter is specified.
Default: ``torch.nn.MSELoss()``.
likelihood
One of Darts' :meth:`Likelihood <darts.utils.likelihood_models.Likelihood>` models to be used for
probabilistic forecasts. Default: ``None``.
torch_metrics
A torch metric or a ``MetricCollection`` used for evaluation. A full list of available metrics can be found
at https://torchmetrics.readthedocs.io/en/latest/. Default: ``None``.
optimizer_cls
The PyTorch optimizer class to be used. Default: ``torch.optim.Adam``.
optimizer_kwargs
Optionally, some keyword arguments for the PyTorch optimizer (e.g., ``{'lr': 1e-3}``
for specifying a learning rate). Otherwise the default values of the selected ``optimizer_cls``
will be used. Default: ``None``.
lr_scheduler_cls
Optionally, the PyTorch learning rate scheduler class to be used. Specifying ``None`` corresponds
to using a constant learning rate. Default: ``None``.
lr_scheduler_kwargs
Optionally, some keyword arguments for the PyTorch learning rate scheduler. Default: ``None``.
use_reversible_instance_norm
Whether to use reversible instance normalization `RINorm` against distribution shift as shown in [1]_.
It is only applied to the features of the target series and not the covariates.
batch_size
Number of time series (input and output sequences) used in each training pass. Default: ``32``.
n_epochs
Number of epochs over which to train the model. Default: ``100``.
model_name
Name of the model. Used for creating checkpoints and saving tensorboard data. If not specified,
defaults to the following string ``"YYYY-mm-dd_HH_MM_SS_torch_model_run_PID"``, where the initial part
of the name is formatted with the local date and time, while PID is the processed ID (preventing models
spawned at the same time by different processes to share the same model_name). E.g.,
``"2021-06-14_09_53_32_torch_model_run_44607"``.
work_dir
Path of the working directory, where to save checkpoints and Tensorboard summaries.
Default: current working directory.
log_tensorboard
If set, use Tensorboard to log the different parameters. The logs will be located in:
``"{work_dir}/darts_logs/{model_name}/logs/"``. Default: ``False``.
nr_epochs_val_period
Number of epochs to wait before evaluating the validation loss (if a validation
``TimeSeries`` is passed to the :func:`fit()` method). Default: ``1``.
force_reset
If set to ``True``, any previously-existing model with the same name will be reset (all checkpoints will
be discarded). Default: ``False``.
save_checkpoints
Whether to automatically save the untrained model and checkpoints from training.
To load the model from checkpoint, call :func:`MyModelClass.load_from_checkpoint()`, where
:class:`MyModelClass` is the :class:`TorchForecastingModel` class that was used (such as :class:`TFTModel`,
:class:`NBEATSModel`, etc.). If set to ``False``, the model can still be manually saved using
:func:`save()` and loaded using :func:`load()`. Default: ``False``.
add_encoders
A large number of past and future covariates can be automatically generated with `add_encoders`.
This can be done by adding multiple pre-defined index encoders and/or custom user-made functions that
will be used as index encoders. Additionally, a transformer such as Darts' :class:`Scaler` can be added to
transform the generated covariates. This happens all under one hood and only needs to be specified at
model creation.
Read :meth:`SequentialEncoder <darts.dataprocessing.encoders.SequentialEncoder>` to find out more about
``add_encoders``. Default: ``None``. An example showing some of ``add_encoders`` features:
.. highlight:: python
.. code-block:: python
def encode_year(idx):
return (idx.year - 1950) / 50
add_encoders={
'cyclic': {'future': ['month']},
'datetime_attribute': {'future': ['hour', 'dayofweek']},
'position': {'past': ['relative'], 'future': ['relative']},
'custom': {'past': [encode_year]},
'transformer': Scaler(),
'tz': 'CET'
}
..
random_state
Control the randomness of the weights initialization. Check this
`link <https://scikit-learn.org/stable/glossary.html#term-random_state>`_ for more details.
Default: ``None``.
pl_trainer_kwargs
By default :class:`TorchForecastingModel` creates a PyTorch Lightning Trainer with several useful presets
that performs the training, validation and prediction processes. These presets include automatic
checkpointing, tensorboard logging, setting the torch device and more.
With ``pl_trainer_kwargs`` you can add additional kwargs to instantiate the PyTorch Lightning trainer
object. Check the `PL Trainer documentation
<https://pytorch-lightning.readthedocs.io/en/stable/common/trainer.html>`_ for more information about the
supported kwargs. Default: ``None``.
Running on GPU(s) is also possible using ``pl_trainer_kwargs`` by specifying keys ``"accelerator",
"devices", and "auto_select_gpus"``. Some examples for setting the devices inside the ``pl_trainer_kwargs``
dict:
- ``{"accelerator": "cpu"}`` for CPU,
- ``{"accelerator": "gpu", "devices": [i]}`` to use only GPU ``i`` (``i`` must be an integer),
- ``{"accelerator": "gpu", "devices": -1, "auto_select_gpus": True}`` to use all available GPUS.
For more info, see here:
https://pytorch-lightning.readthedocs.io/en/stable/common/trainer.html#trainer-flags , and
https://pytorch-lightning.readthedocs.io/en/stable/accelerators/gpu_basic.html#train-on-multiple-gpus
With parameter ``"callbacks"`` you can add custom or PyTorch-Lightning built-in callbacks to Darts'
:class:`TorchForecastingModel`. Below is an example for adding EarlyStopping to the training process.
The model will stop training early if the validation loss `val_loss` does not improve beyond
specifications. For more information on callbacks, visit:
`PyTorch Lightning Callbacks
<https://pytorch-lightning.readthedocs.io/en/stable/extensions/callbacks.html>`_
.. highlight:: python
.. code-block:: python
from pytorch_lightning.callbacks.early_stopping import EarlyStopping
# stop training when validation loss does not decrease more than 0.05 (`min_delta`) over
# a period of 5 epochs (`patience`)
my_stopper = EarlyStopping(
monitor="val_loss",
patience=5,
min_delta=0.05,
mode='min',
)
pl_trainer_kwargs={"callbacks": [my_stopper]}
..
Note that you can also use a custom PyTorch Lightning Trainer for training and prediction with optional
parameter ``trainer`` in :func:`fit()` and :func:`predict()`.
show_warnings
whether to show warnings raised from PyTorch Lightning. Useful to detect potential issues of
your forecasting use case. Default: ``False``.
References
----------
.. [1] T. Kim et al. "Reversible Instance Normalization for Accurate Time-Series Forecasting against
Distribution Shift", https://openreview.net/forum?id=cGDAkQo1C0p
Examples
--------
>>> from darts.datasets import WeatherDataset
>>> from darts.models import BlockRNNModel
>>> series = WeatherDataset().load()
>>> # predicting atmospheric pressure
>>> target = series['p (mbar)'][:100]
>>> # optionally, use past observed rainfall (pretending to be unknown beyond index 100)
>>> past_cov = series['rain (mm)'][:100]
>>> # predict 6 pressure values using the 12 past values of pressure and rainfall, as well as the 6 temperature
>>> model = BlockRNNModel(
>>> input_chunk_length=12,
>>> output_chunk_length=6,
>>> n_rnn_layers=2,
>>> n_epochs=50,
>>> )
>>> model.fit(target, past_covariates=past_cov)
>>> pred = model.predict(6)
>>> pred.values()
array([[4.97979827],
[3.9707572 ],
[5.27869295],
[5.19697244],
[5.28424783],
[5.22497681]])
.. note::
`RNN example notebook <https://unit8co.github.io/darts/examples/04-RNN-examples.html>`_ presents techniques
that can be used to improve the forecasts quality compared to this simple usage example.
"""
super().__init__(**self._extract_torch_model_params(**self.model_params))
# extract pytorch lightning module kwargs
self.pl_module_params = self._extract_pl_module_params(**self.model_params)
# check we got right model type specified:
if model not in ["RNN", "LSTM", "GRU"]:
if not inspect.isclass(model) or not issubclass(
model, CustomBlockRNNModule
):
raise_log(
ValueError(
"`model` is not a valid RNN model. Please specify 'RNN', 'LSTM', 'GRU', or give a subclass "
"(not an instance) of darts.models.forecasting.rnn_model.CustomBlockRNNModule."
),
logger=logger,
)
self.rnn_type_or_module = model
self.hidden_fc_sizes = hidden_fc_sizes
self.hidden_dim = hidden_dim
self.n_rnn_layers = n_rnn_layers
self.dropout = dropout
@property
def supports_multivariate(self) -> bool:
return True
def _create_model(self, train_sample: Tuple[torch.Tensor]) -> torch.nn.Module:
# samples are made of (past_target, past_covariates, future_target)
input_dim = train_sample[0].shape[1] + (
train_sample[1].shape[1] if train_sample[1] is not None else 0
)
output_dim = train_sample[-1].shape[1]
nr_params = 1 if self.likelihood is None else self.likelihood.num_parameters
hidden_fc_sizes = [] if self.hidden_fc_sizes is None else self.hidden_fc_sizes
kwargs = {}
if isinstance(self.rnn_type_or_module, str):
model_cls = _BlockRNNModule
kwargs["name"] = self.rnn_type_or_module
else:
model_cls = self.rnn_type_or_module
return model_cls(
input_size=input_dim,
target_size=output_dim,
nr_params=nr_params,
hidden_dim=self.hidden_dim,
num_layers=self.n_rnn_layers,
num_layers_out_fc=hidden_fc_sizes,
dropout=self.dropout,
**self.pl_module_params,
**kwargs,
)