DeepAbstractNet¶
- class DeepAbstractNet[source]¶
Bases:
abc.ABCInterface for etna native deep models.
- Inherited-members
Methods
make_samples(df, encoder_length, decoder_length)Make samples from input slice of TSDataset.
step(batch, *args, **kwargs)Make batch step.
- abstract make_samples(df: pandas.core.frame.DataFrame, encoder_length: int, decoder_length: int) Iterable[dict][source]¶
Make samples from input slice of TSDataset.
- Parameters
df (pandas.core.frame.DataFrame) – slice is per-segment Dataframes
encoder_length (int) – encoder_length
decoder_length (int) – decoder_length
- Returns
samples of input slices
- Return type
Iterable[dict]
- abstract step(batch: dict, *args, **kwargs) Tuple[torch.Tensor, torch.Tensor, torch.Tensor][source]¶
Make batch step.
- Parameters
batch (dict) – Batch with data to make inference on.
- Returns
loss, true_target, prediction_target
- Return type
Tuple[torch.Tensor, torch.Tensor, torch.Tensor]