EventStream.transformer.fine_tuning_model module¶

A model for fine-tuning on classification tasks.

class EventStream.transformer.fine_tuning_model.ESTForStreamClassification(config: StructuredTransformerConfig)[source]¶

Bases: StructuredTransformerPreTrainedModel

A model for fine-tuning on classification tasks.

Parameters:¶

config: StructuredTransformerConfig¶

The model configuration class to use. This must contain the relevant fine-tuning task information (e.g., num_labels, finetuning_task, pooling_method, and id2label).

forward(batch: PytorchBatch, **kwargs) → StreamClassificationModelOutput[source]¶

Runs the forward pass through the fine-tuning label prediction.

Parameters:¶

batch: PytorchBatch¶

The batch of data to model.

Returns:¶

A StreamClassificationModelOutput object capturing loss, predictions, and labels for the fine-tuning task in question.