base_model#

Timeseries models share a number of common characteristics. This module implements these in a common base class.

Functions

_concatenate_output(output)

Concatenate multiple batches of output dictionary.

_torch_cat_na(x)

Concatenate tensor along dim=0 and add nans along dim=1 if necessary.

Classes

AutoRegressiveBaseModel([log_interval, ...])

Model with additional methods for autoregressive models.

AutoRegressiveBaseModelWithCovariates([...])

Model with additional methods for autoregressive models with covariates.

BaseModel([log_interval, log_val_interval, ...])

BaseModel from which new timeseries models should inherit from.

BaseModelWithCovariates([log_interval, ...])

Model with additional methods using covariates.

PredictCallback([mode, return_index, ...])

Internally used callback to capture predictions and optionally write them to disk.

PredictTuple

alias of prediction

Prediction([output, x, index, ...])

Create new instance of prediction(output, x, index, decoder_lengths, y)