architecture

tiatoolbox . models . architecture

Defines a set of models to be used within tiatoolbox.

Functions

fetch_pretrained_weights	Get the pretrained model information from yml file. :param model_name: Refer to ::py::meth:get_pretrained_model for all supported model names. :type model_name: str :param save_path: Path to save the weight of the corresponding model_name. :type save_path: str :param overwrite: To always overwriting downloaded weights. :type overwrite: bool.
get_pretrained_model	Load a predefined PyTorch model with the appropriate pretrained weights.

tiatoolbox.models.architecture.hovernet
tiatoolbox.models.architecture.hovernetplus
tiatoolbox.models.architecture.idars	Defines CNNs as used in IDaRS for prediction of molecular pathways and mutations.
tiatoolbox.models.architecture.unet	Defines a set of ResNet variants to be used within tiatoolbox.
tiatoolbox.models.architecture.utils	Defines utlity layers and operators for models in tiatoolbox.
tiatoolbox.models.architecture.vanilla	Defines vanilla CNNs with torch backbones, mainly for patch classification.