finetuner package

Subpackages

Submodules

Module contents

finetuner.fit(model: AnyDNN, train_data: DocumentSequence, eval_data: Optional[DocumentSequence] = None, epochs: int = 10, batch_size: int = 256, loss: Union[str, AnyDNN] = 'SiameseLoss', optimizer: Optional[AnyOptimizer] = None, learning_rate: float = 0.001, device: str = 'cpu', preprocess_fn: Optional[PreprocFnType] = None, collate_fn: Optional[CollateFnType] = None, num_items_per_class: Optional[int] = None) Tuple[AnyDNN, Summary][source]
finetuner.fit(model: AnyDNN, train_data: DocumentSequence, eval_data: Optional[DocumentSequence] = None, epochs: int = 10, batch_size: int = 256, loss: Union[str, AnyDNN] = 'SiameseLoss', optimizer: Optional[AnyOptimizer] = None, learning_rate: float = 0.001, device: str = 'cpu', preprocess_fn: Optional[PreprocFnType] = None, collate_fn: Optional[CollateFnType] = None, num_items_per_class: Optional[int] = None, to_embedding_model: bool = True, input_size: Optional[Tuple[int, ...]] = None, input_dtype: str = 'float32', layer_name: Optional[str] = None, output_dim: Optional[int] = None, freeze: bool = False) Tuple[AnyDNN, Summary]
finetuner.fit(model: AnyDNN, train_data: DocumentSequence, eval_data: Optional[DocumentSequence] = None, epochs: int = 10, batch_size: int = 256, loss: Union[str, AnyDNN] = 'SiameseLoss', optimizer: Optional[AnyOptimizer] = None, learning_rate: float = 0.001, device: str = 'cpu', preprocess_fn: Optional[PreprocFnType] = None, collate_fn: Optional[CollateFnType] = None, num_items_per_class: Optional[int] = None, interactive: bool = True, clear_labels_on_start: bool = False, port_expose: Optional[int] = None, runtime_backend: str = 'thread') Tuple[AnyDNN, None]
finetuner.fit(model: AnyDNN, train_data: DocumentSequence, eval_data: Optional[DocumentSequence] = None, epochs: int = 10, batch_size: int = 256, loss: Union[str, AnyDNN] = 'SiameseLoss', optimizer: Optional[AnyOptimizer] = None, learning_rate: float = 0.001, device: str = 'cpu', preprocess_fn: Optional[PreprocFnType] = None, collate_fn: Optional[CollateFnType] = None, num_items_per_class: Optional[int] = None, interactive: bool = True, clear_labels_on_start: bool = False, port_expose: Optional[int] = None, runtime_backend: str = 'thread', to_embedding_model: bool = True, input_size: Optional[Tuple[int, ...]] = None, input_dtype: str = 'float32', layer_name: Optional[str] = None, output_dim: Optional[int] = None, freeze: bool = False) Tuple[AnyDNN, None]
Return type

Tuple[ForwardRef, Optional[ForwardRef]]