pyiqa

Submodules

• pyiqa.api_helpers
• pyiqa.archs
• pyiqa.data
• pyiqa.default_model_configs
• pyiqa.losses
• pyiqa.matlab_utils
• pyiqa.metrics
• pyiqa.models
• pyiqa.pyiqa_cmd
• pyiqa.test
• pyiqa.train
• pyiqa.train_nsplits
• pyiqa.utils

Package Contents

pyiqa.create_metric(metric_name, as_loss=False, device=None, **kwargs)

pyiqa.list_models(metric_mode=None, filter='', exclude_filters='')

Return list of available model names, sorted alphabetically :param filter: :type filter: str :param exclude_filters: :type exclude_filters: str or list[str]

Example

model_list(’ssim’) – returns all models including ‘ssim’

pyiqa.get_dataset_info(dataset_name=None)

pyiqa.load_dataset(name: str, data_root: str = './datasets', force_download: bool = False, dataset_opts: Dict[str, Any] | None = None, **kwargs: Any) → Any

Load a dataset from specified location, downloading if necessary.

Parameters:

• name (str) – Name of the dataset to load

• data_root (str) – Root directory for dataset storage

• force_download (bool) – Whether to force download even if files exist

• dataset_opts (Optional[Dict[str, Any]]) – Additional dataset options

• **kwargs (Any) – Additional arguments passed to dataset configuration

Returns:

Dataset object

Return type:

Any