Inception time

polars_ts.classification.inception_time

InceptionTime classifier for time series.

Implements the InceptionTime architecture (Fawaz et al., 2020) — an ensemble of Inception modules with residual connections for time series classification.

References

Fawaz et al. (2020). InceptionTime: Finding AlexNet for Time Series Classification. Data Mining and Knowledge Discovery.

_InceptionModule

Bases: Module

Single Inception module with multi-scale 1D convolutions.

_InceptionBlock

Bases: Module

Inception block with residual connection.

_InceptionTimeNet

Bases: Module

Full InceptionTime network.

InceptionTimeClassifier

InceptionTime time series classifier.

Parameters

n_filters Number of filters per Inception module branch. max_epochs Maximum training epochs. lr Learning rate. batch_size Training batch size. id_col, target_col, time_col Column names.

fit(df, *, label_col='label')

Train on labeled time series data.

predict(df)

Predict class labels for test series.