brails.processors.cons_type_classifier.ConstructionTypeClassifier module

Class object to use or create construction type classification models.

class brails.processors.cons_type_classifier.ConstructionTypeClassifier.ConsTypeClassifier(model_path: str | None = None)

Bases: ImageClassifier

Class for facilitating classification of different construction types.

ConsTypeClassifier is a specialized class for classifying different construction types (e.g., MAB, MAS, RCC, STL, WOD). It allows for making predictions using a pre-trained model and retraining the pre-trained model on new data. Inherits from the ImageClassifier class.

This class loads a default pre-trained model if a custom model path is not provided during initialization. It can also be retrained using new datasets.

predict(data_dir: str) → None

Perform construction type predictions on images in the specified path.

Args__

data_dir (str): Path to the directory containing images to be

classified.

retrain(data_dir: str, batch_size: int = 8, nepochs: int = 100, plot_loss: bool = True) → None

Retrain the construction type classifier on new data.

Args__

data_dir (str): Path to the directory containing training data. batch_size (int, optional): Batch size for training. Default is 8. nepochs (int, optional): Number of epochs for training. Default is

plot_loss (bool, optional): Whether to plot the loss during

training. Default is True.