PyTorch type#

class flytekit.extras.pytorch.PyTorchCheckpoint(module=None, hyperparameters=None, optimizer=None)#

This class is helpful to save a checkpoint.

Parameters:

  • module (Module | None)

  • hyperparameters (Dict[str, Any] | NamedTuple | IsDataclass | None)

  • optimizer (Optimizer | None)

class flytekit.extras.pytorch.PyTorchCheckpointTransformer#

TypeTransformer that supports serializing and deserializing checkpoint.

get_literal_type(t)#

Converts the python type to a Flyte LiteralType

Parameters:

t (Type[PyTorchCheckpoint])

Return type:

LiteralType

guess_python_type(literal_type)#

Converts the Flyte LiteralType to a python object type.

Parameters:

literal_type (LiteralType)

Return type:

Type[PyTorchCheckpoint]

to_literal(ctx, python_val, python_type, expected)#

Converts a given python_val to a Flyte Literal, assuming the given python_val matches the declared python_type. Implementers should refrain from using type(python_val) instead rely on the passed in python_type. If these do not match (or are not allowed) the Transformer implementer should raise an AssertionError, clearly stating what was the mismatch :param ctx: A FlyteContext, useful in accessing the filesystem and other attributes :param python_val: The actual value to be transformed :param python_type: The assumed type of the value (this matches the declared type on the function) :param expected: Expected Literal Type

Parameters:
Return type:

Literal

to_python_value(ctx, lv, expected_python_type)#

Converts the given Literal to a Python Type. If the conversion cannot be done an AssertionError should be raised :param ctx: FlyteContext :param lv: The received literal Value :param expected_python_type: Expected native python type that should be returned

Parameters:
Return type:

PyTorchCheckpoint

class flytekit.extras.pytorch.PyTorchModuleTransformer#
guess_python_type(literal_type)#

Converts the Flyte LiteralType to a python object type.

Parameters:

literal_type (LiteralType)

Return type:

Type[Module]

class flytekit.extras.pytorch.PyTorchTensorTransformer#
guess_python_type(literal_type)#

Converts the Flyte LiteralType to a python object type.

Parameters:

literal_type (LiteralType)

Return type:

Type[Tensor]