classy.pl_modules.hf.classification

Classes

HFQAPLModule

class HFQAPLModule()

##

Simple Mixin to model the prediction behavior of a classy.pl_modules.base.ClassyPLModule.

__init__

def __init__(
    transformer_model: str,
    optim_conf: omegaconf.dictconfig.DictConfig,
    additional_special_tokens: Optional[List[str]] = None,
)

##

test_step

def test_step(
    self,
    batch: dict,
    batch_idx: int,
) ‑> None

##

training_step

def training_step(
    self,
    batch: dict,
    batch_idx: int,
) ‑> torch.Tensor

##

validation_step

def validation_step(
    self,
    batch: dict,
    batch_idx: int,
) ‑> None

##

HFSentencePairPLModule

class HFSentencePairPLModule()

##

Simple Mixin to model the prediction behavior of a classy.pl_modules.base.ClassyPLModule.

__init__

def __init__(
    transformer_model: str,
    vocabulary: Vocabulary,
    optim_conf: omegaconf.dictconfig.DictConfig,
    additional_special_tokens: Optional[List[str]] = None,
)

##

HFSequenceCommonPLModule

class HFSequenceCommonPLModule()

##

Simple Mixin to model the prediction behavior of a classy.pl_modules.base.ClassyPLModule.

Subclasses (2)

• HFSentencePairPLModule
• HFSequencePLModule

__init__

def __init__(
    transformer_model: str,
    vocabulary: Vocabulary,
    optim_conf: omegaconf.dictconfig.DictConfig,
    additional_special_tokens: Optional[List[str]] = None,
)

##

test_step

def test_step(
    self,
    batch: dict,
    batch_idx: int,
) ‑> None

##

training_step

def training_step(
    self,
    batch: dict,
    batch_idx: int,
) ‑> torch.Tensor

##

validation_step

def validation_step(
    self,
    batch: dict,
    batch_idx: int,
) ‑> None

##

HFSequencePLModule

class HFSequencePLModule()

##

Simple Mixin to model the prediction behavior of a classy.pl_modules.base.ClassyPLModule.

__init__

def __init__(
    transformer_model: str,
    vocabulary: Vocabulary,
    optim_conf: omegaconf.dictconfig.DictConfig,
    additional_special_tokens: Optional[List[str]] = None,
)

##

HFTokensPLModule

class HFTokensPLModule()

##

Simple Mixin to model the prediction behavior of a classy.pl_modules.base.ClassyPLModule.

__init__

def __init__(
    transformer_model: str,
    use_last_n_layers: int,
    fine_tune: bool,
    vocabulary: Vocabulary,
    optim_conf: omegaconf.dictconfig.DictConfig,
    additional_special_tokens: Optional[List[str]] = None,
)

##

test_step

def test_step(
    self,
    batch: dict,
    batch_idx: int,
) ‑> None

##

training_step

def training_step(
    self,
    batch: dict,
    batch_idx: int,
) ‑> torch.Tensor

##

validation_step

def validation_step(
    self,
    batch: dict,
    batch_idx: int,
) ‑> None

##