Model Training

Model training module for training SpaCy NER models.

This module provides functionality to train and fine-tune SpaCy models using generated or custom training data.

class spacylize.trainer.ModelTrainer(train_data, base_model, output_model, n_iter, dropout)

Bases: object

Trainer for SpaCy NER models.

Trains or fine-tunes a SpaCy model using provided training data with configurable hyperparameters.

Parameters:

• train_data (Path)

• base_model (str)

• output_model (Path)

• n_iter (int)

• dropout (float)

train_data

Path to the training dataset (.spacy file).

base_model

Name of the base SpaCy model to train/fine-tune.

output_model

Path where the trained model will be saved.

n_iter

Number of training iterations.

dropout

Dropout rate during training for regularization.

Note

This class is not yet fully implemented.

__init__(train_data, base_model, output_model, n_iter, dropout)

Initialize the ModelTrainer.

Parameters:

• train_data (Path) – Path to the training data file (.spacy).

• base_model (str) – Base SpaCy model name (e.g., ‘en_core_web_sm’).

• output_model (Path) – Path to save the trained model.

• n_iter (int) – Number of training iterations.

• dropout (float) – Dropout rate (0.0-1.0) for regularization.

run()

Run the training process.

Note

This method is not yet implemented.