PatchTSModel¶

class PatchTSModel(decoder_length: int, encoder_length: int, patch_len: int = 4, stride: int = 1, num_layers: int = 3, hidden_size: int = 128, feedforward_size: int = 256, nhead: int = 16, lr: float = 0.001, loss: Optional[torch.nn.modules.module.Module] = None, train_batch_size: int = 128, test_batch_size: int = 128, optimizer_params: Optional[dict] = None, trainer_params: Optional[dict] = None, train_dataloader_params: Optional[dict] = None, test_dataloader_params: Optional[dict] = None, val_dataloader_params: Optional[dict] = None, split_params: Optional[dict] = None)[source]¶

Bases: etna.models.base.DeepBaseModel

PatchTS model using PyTorch layers.

Init PatchTS model.

Parameters

encoder_length (int) – encoder length
decoder_length (int) –
patch_len (int) –
stride (int) –
num_layers (int) –
hidden_size (int) –
feedforward_size (int) –
nhead (int) –
lr (float) –
loss (Optional[torch.nn.Module]) –
train_batch_size (int) –
test_batch_size (int) –
optimizer_params (Optional[dict]) –
trainer_params (Optional[dict]) –
train_dataloader_params (Optional[dict]) –
test_dataloader_params (Optional[dict]) –
val_dataloader_params (Optional[dict]) –
split_params (Optional[dict]) –

decoder_length:

decoder length

patch_len:

size of patch

stride:

step of patch

num_layers:

number of layers

hidden_size:

size of the hidden state

feedforward_size:

size of feedforward layers in transformer

nhead:

number of transformer heads

lr:

learning rate

loss:

loss function, MSELoss by default

train_batch_size:

batch size for training

test_batch_size:

batch size for testing

optimizer_params:

parameters for optimizer for Adam optimizer (api reference torch.optim.Adam)

trainer_params:

Pytorch ligthning trainer parameters (api reference pytorch_lightning.trainer.trainer.Trainer)

train_dataloader_params:

parameters for train dataloader like sampler for example (api reference torch.utils.data.DataLoader)

test_dataloader_params:

parameters for test dataloader

val_dataloader_params:

parameters for validation dataloader

split_params:

dictionary with parameters for torch.utils.data.random_split() for train-test splitting

train_size: (float) value from 0 to 1 - fraction of samples to use for training
generator: (Optional[torch.Generator]) - generator for reproducibile train-test splitting
torch_dataset_size: (Optional[int]) - number of samples in dataset, in case of dataset not implementing __len__

Inherited-members

Parameters

decoder_length (int) –
encoder_length (int) –
patch_len (int) –
stride (int) –
num_layers (int) –
hidden_size (int) –
feedforward_size (int) –
nhead (int) –
lr (float) –
loss (Optional[torch.nn.Module]) –
train_batch_size (int) –
test_batch_size (int) –
optimizer_params (Optional[dict]) –
trainer_params (Optional[dict]) –
train_dataloader_params (Optional[dict]) –
test_dataloader_params (Optional[dict]) –
val_dataloader_params (Optional[dict]) –
split_params (Optional[dict]) –

Methods

`fit`(ts)	Fit model.
`forecast`(ts, prediction_size[, ...])	Make predictions.
`get_model`()	Get model.
`load`(path)	Load an object.
`params_to_tune`()	Get default grid for tuning hyperparameters.
`predict`(ts, prediction_size[, return_components])	Make predictions.
`raw_fit`(torch_dataset)	Fit model on torch like Dataset.
`raw_predict`(torch_dataset)	Make inference on torch like Dataset.
`save`(path)	Save the object.
`set_params`(**params)	Return new object instance with modified parameters.
`to_dict`()	Collect all information about etna object in dict.

Attributes

context_size

Context size of the model.

fit(ts: etna.datasets.tsdataset.TSDataset) → etna.models.base.DeepBaseModel¶

Fit model.

Parameters: ts (etna.datasets.tsdataset.TSDataset) – TSDataset with features
Returns: Model after fit
Return type: etna.models.base.DeepBaseModel

forecast(ts: etna.datasets.tsdataset.TSDataset, prediction_size: int, return_components: bool = False) → etna.datasets.tsdataset.TSDataset¶

Make predictions.

This method will make autoregressive predictions.

Parameters

ts (etna.datasets.tsdataset.TSDataset) – Dataset with features and expected decoder length for context
prediction_size (int) – Number of last timestamps to leave after making prediction. Previous timestamps will be used as a context.
return_components (bool) – If True additionally returns forecast components

Returns

Dataset with predictions

Return type

etna.datasets.tsdataset.TSDataset

get_model() → etna.models.base.DeepBaseNet¶

Get model.

Returns: Torch Module
Return type: etna.models.base.DeepBaseNet

classmethod load(path: pathlib.Path) → typing_extensions.Self¶

Load an object.

Warning

This method uses dill module which is not secure. It is possible to construct malicious data which will execute arbitrary code during loading. Never load data that could have come from an untrusted source, or that could have been tampered with.

Parameters: path (pathlib.Path) – Path to load object from.
Returns: Loaded object.
Return type: typing_extensions.Self

params_to_tune() → Dict[str, etna.distributions.distributions.BaseDistribution][source]¶

Get default grid for tuning hyperparameters.

This grid tunes parameters: num_layers, hidden_size, lr, encoder_length. Other parameters are expected to be set by the user.

Returns: Grid to tune.
Return type: Dict[str, etna.distributions.distributions.BaseDistribution]

predict(ts: etna.datasets.tsdataset.TSDataset, prediction_size: int, return_components: bool = False) → etna.datasets.tsdataset.TSDataset¶

Make predictions.

This method will make predictions using true values instead of predicted on a previous step. It can be useful for making in-sample forecasts.

Parameters

ts (etna.datasets.tsdataset.TSDataset) – Dataset with features and expected decoder length for context
prediction_size (int) – Number of last timestamps to leave after making prediction. Previous timestamps will be used as a context.
return_components (bool) – If True additionally returns prediction components

Returns

Dataset with predictions

Return type

etna.datasets.tsdataset.TSDataset

raw_fit(torch_dataset: torch.utils.data.dataset.Dataset) → etna.models.base.DeepBaseModel¶

Fit model on torch like Dataset.

Parameters: torch_dataset (torch.utils.data.dataset.Dataset) – Torch like dataset for model fit
Returns: Model after fit
Return type: etna.models.base.DeepBaseModel

raw_predict(torch_dataset: torch.utils.data.dataset.Dataset) → Dict[Tuple[str, str], numpy.ndarray]¶

Make inference on torch like Dataset.

Parameters: torch_dataset (torch.utils.data.dataset.Dataset) – Torch like dataset for model inference
Returns: Dictionary with predictions
Return type: Dict[Tuple[str, str], numpy.ndarray]

save(path: pathlib.Path)¶

Save the object.

Parameters: path (pathlib.Path) – Path to save object to.

set_params(**params: dict) → etna.core.mixins.TMixin¶

Return new object instance with modified parameters.

Method also allows to change parameters of nested objects within the current object. For example, it is possible to change parameters of a model in a Pipeline.

Nested parameters are expected to be in a <component_1>.<...>.<parameter> form, where components are separated by a dot.

Parameters

**params – Estimator parameters
self (etna.core.mixins.TMixin) –
params (dict) –

Returns

New instance with changed parameters

Return type

etna.core.mixins.TMixin

Examples

>>> from etna.pipeline import Pipeline
>>> from etna.models import NaiveModel
>>> from etna.transforms import AddConstTransform
>>> model = model=NaiveModel(lag=1)
>>> transforms = [AddConstTransform(in_column="target", value=1)]
>>> pipeline = Pipeline(model, transforms=transforms, horizon=3)
>>> pipeline.set_params(**{"model.lag": 3, "transforms.0.value": 2})
Pipeline(model = NaiveModel(lag = 3, ), transforms = [AddConstTransform(in_column = 'target', value = 2, inplace = True, out_column = None, )], horizon = 3, )

to_dict()¶: Collect all information about etna object in dict.

property context_size: int¶: Context size of the model.