DeepAR native implementation

CHANGELOG.md

@@ -8,9 +8,12 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
 ## Unreleased
 ### Added
 - Add params_to_tune for DeepStateModel ([#115](https://github.com/etna-team/etna/issues/115))
+- 
+- 
 - Handle new functionality for prediction intervals in the `plot_forecast` ([#130](https://github.com/etna-team/etna/pull/130))  
 - Add `get_historical_forecasts` to pipelines for forecast estimation at each fold on the historical dataset ([#143](https://github.com/etna-team/etna/pull/143))
-- 
+- Add DeepARNativeModel ([#114](https://github.com/etna-team/etna/pull/114))
+-
 - 
 - 
 - 

docs/source/api_reference/models.rst

@@ -81,6 +81,7 @@ Native neural network models:
    nn.NBeatsGenericModel
    nn.NBeatsInterpretableModel
    nn.PatchTSModel
+   nn.DeepARNativeModel
 
 Utilities for :py:class:`~etna.models.nn.deepstate.deepstate.DeepStateModel`
 

etna/models/nn/__init__.py

@@ -2,6 +2,7 @@
 
 if SETTINGS.torch_required:
     from etna.models.nn.deepar import DeepARModel
+    from etna.models.nn.deepar_native import DeepARNativeModel
     from etna.models.nn.deepstate.deepstate import DeepStateModel
     from etna.models.nn.mlp import MLPModel
     from etna.models.nn.nbeats import NBeatsGenericModel

etna/models/nn/deepar.py

@@ -4,6 +4,7 @@
 from typing import Sequence
 
 import pandas as pd
+from deprecated import deprecated
 
 from etna import SETTINGS
 from etna.datasets.tsdataset import TSDataset
@@ -27,6 +28,7 @@
     from pytorch_lightning import Trainer
 
 
+@deprecated(reason="DeepARModel is deprecated. Use DeepARNativeModel instead.", version="3.0")
 class DeepARModel(
     _DeepCopyMixin, PytorchForecastingMixin, SavePytorchForecastingMixin, PredictionIntervalContextRequiredAbstractModel
 ):

etna/models/nn/deepar_native/__init__.py

@@ -0,0 +1,4 @@
+from etna import SETTINGS
+
+if SETTINGS.torch_required:
+    from etna.models.nn.deepar_native.deepar import DeepARNativeModel