Module tripleblind.tasks

These Tasks represent targeted jobs and preprocessing for specific things. The general hierarchy is:

TrainingTask
    TrainTabularNet
    TrainImageNet

InferenceTask
    InferTabularNet
    InferImage

Classes

class TrainTabularNet (job_name: str, network: Asset, columns: int, datasets: List[Asset], epochs: int, inference_type: str, loss_function: str, optimizer: str, optimizer_params: dict)

A single operation involving an algorithm or training protocol, some form of data (either a data asset or an input file) and any parameters used by the algorithm/protocol.

Ancestors

Inherited members

class TrainingTask (job_name: str, training_model: Asset, dataset: List[Asset], params: dict)

A single operation involving an algorithm or training protocol, some form of data (either a data asset or an input file) and any parameters used by the algorithm/protocol.

Ancestors

Subclasses

Class variables

var trained_network : Asset

After training, the Asset representing the trained neural network

var training_model : Asset

After training this holds the trained neural network

Methods

def train(self, quiet: Optional[bool] = False) -> bool

Inherited members