detoxai.methods.other

Submodules

detoxai.methods.other.fine_tune submodule

class detoxai.methods.other.fine_tune.FineTune(model: LightningModule, experiment_name: str, device: str, **kwargs)[source]

Bases: ModelCorrectionMethod

This is kind-of a dummy correction method that is a baseline in form of further fine tuning of the model

apply_model_correction(dataloader: DataLoader, logger: object | bool = False, fine_tune_epochs: int = 1, lr: float = 0.0001, **kwargs) → None[source]
Parameters:

  • dataloader – torch.utils.data.DataLoader:

  • logger – object | bool: (Default value = False)

  • fine_tune_epochs – int: (Default value = 1)

  • lr – float: (Default value = 1e-4)

  • **kwargs

Returns:

Module contents