hezar.data.datasets.image_captioning_dataset module

class hezar.data.datasets.image_captioning_dataset.ImageCaptioningDataset(config: ImageCaptioningDatasetConfig, split=None, **kwargs)

Bases: Dataset

required_backends: List[str | Backends] = [Backends.SCIKIT]

class hezar.data.datasets.image_captioning_dataset.ImageCaptioningDatasetConfig(task: TaskType = TaskType.IMAGE2TEXT, path: str | None = None, tokenizer_path: str | None = None, text_column: str = 'label', max_length: int | None = None, test_split_size: float = 0.2, image_processor_config: ImageProcessorConfig | None = None)

Bases: DatasetConfig

Configuration class for image captioning datasets.

Parameters:

• path (str) – Path to the dataset.

• tokenizer_path (str) – Path to the tokenizer file.

• text_column (str) – Column name for text in the dataset.

• images_paths_column (str) – Column name for image paths in the dataset.

• max_length (int) – Maximum length of text.

• test_split_size (float) – Size of the test split.

• image_processor_config (ImageProcessorConfig) – Configuration for image processing.

image_processor_config: ImageProcessorConfig = None

images_paths_column = 'image_path'

max_length: int = None

name: str = 'image_captioning'

path: str = None

task: TaskType = 'image2text'

test_split_size: float = 0.2

text_column: str = 'label'

tokenizer_path: str = None