Skip to main content

classy.pl_modules.hf.generation

Classes​

BartGenerativeModule​

class BartGenerativeModule()

##

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

Subclasses (1)

__init__​

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

load_prediction_params​

def load_prediction_params(
    self,
    prediction_params: Dict[~KT, ~VT],
)
##

GPT2GenerativeModule​

class GPT2GenerativeModule()

##

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

__init__​

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

load_prediction_params​

def load_prediction_params(
    self,
    prediction_params: Dict[~KT, ~VT],
)
##

HFGenerationPLModule​

class HFGenerationPLModule()

##

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

Subclasses (3)

__init__​

def __init__(
    vocabulary: Optional[Vocabulary],
    optim_conf: omegaconf.dictconfig.DictConfig,
)
##

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
##

MBartGenerativeModule​

class MBartGenerativeModule()

##

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

__init__​

def __init__(
    *args,
    **kwargs,
)
##

T5GenerativeModule​

class T5GenerativeModule()

##

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

__init__​

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

load_prediction_params​

def load_prediction_params(
    self,
    prediction_params: Dict[~KT, ~VT],
)
##