Skip to content

GitLab

Unverified Commit 378142af authored by Sylvain Gugger's avatar Sylvain Gugger Committed by GitHub
Browse files

Rename add_start_docstrings_to_callable (#8120)

parent 6241c873
Hide whitespace changes
Inline Side-by-side
Showing with 96 additions and 84 deletions
src/transformers/modeling_tf_gpt2.py
View file @ 378142af
......@@ -27,7 +27,7 @@ from .file_utils import (
ModelOutput,
add_code_sample_docstrings,
add_start_docstrings,
add_start_docstrings_to_callable,
add_start_docstrings_to_model_forward,
replace_return_docstrings,
)
from .modeling_tf_outputs import TFBaseModelOutputWithPast, TFCausalLMOutputWithPast
......@@ -557,7 +557,7 @@ class TFGPT2Model(TFGPT2PreTrainedModel):
super().__init__(config, *inputs, **kwargs)
self.transformer = TFGPT2MainLayer(config, name="transformer")
@add_start_docstrings_to_callable(GPT2_INPUTS_DOCSTRING)
@add_start_docstrings_to_model_forward(GPT2_INPUTS_DOCSTRING)
@add_code_sample_docstrings(
tokenizer_class=_TOKENIZER_FOR_DOC,
checkpoint="gpt2",
......@@ -591,7 +591,7 @@ class TFGPT2LMHeadModel(TFGPT2PreTrainedModel, TFCausalLanguageModelingLoss):
return {"inputs": inputs, "past": past, "use_cache": kwargs["use_cache"]}
@add_start_docstrings_to_callable(GPT2_INPUTS_DOCSTRING)
@add_start_docstrings_to_model_forward(GPT2_INPUTS_DOCSTRING)
@add_code_sample_docstrings(
tokenizer_class=_TOKENIZER_FOR_DOC,
checkpoint="gpt2",
......@@ -687,7 +687,7 @@ class TFGPT2DoubleHeadsModel(TFGPT2PreTrainedModel):
def get_output_embeddings(self):
return self.transformer.wte
@add_start_docstrings_to_callable(GPT2_INPUTS_DOCSTRING)
@add_start_docstrings_to_model_forward(GPT2_INPUTS_DOCSTRING)
@replace_return_docstrings(output_type=TFGPT2DoubleHeadsModelOutput, config_class=_CONFIG_FOR_DOC)
def call(
self,
......
src/transformers/modeling_tf_longformer.py
View file @ 378142af
......@@ -19,7 +19,7 @@ import tensorflow as tf
from transformers.activations_tf import get_tf_activation
from .configuration_longformer import LongformerConfig
from .file_utils import add_code_sample_docstrings, add_start_docstrings, add_start_docstrings_to_callable
from .file_utils import add_code_sample_docstrings, add_start_docstrings, add_start_docstrings_to_model_forward
from .modeling_tf_outputs import (
TFBaseModelOutput,
TFBaseModelOutputWithPooling,
......@@ -1624,7 +1624,7 @@ class TFLongformerModel(TFLongformerPreTrainedModel):
self.longformer = TFLongformerMainLayer(config, name="longformer")
@add_start_docstrings_to_callable(LONGFORMER_INPUTS_DOCSTRING.format("batch_size, sequence_length"))
@add_start_docstrings_to_model_forward(LONGFORMER_INPUTS_DOCSTRING.format("batch_size, sequence_length"))
def call(self, inputs, **kwargs):
outputs = self.longformer(inputs, **kwargs)
......@@ -1648,7 +1648,7 @@ class TFLongformerForMaskedLM(TFLongformerPreTrainedModel, TFMaskedLanguageModel
def get_output_embeddings(self):
return self.lm_head.decoder
@add_start_docstrings_to_callable(LONGFORMER_INPUTS_DOCSTRING.format("batch_size, sequence_length"))
@add_start_docstrings_to_model_forward(LONGFORMER_INPUTS_DOCSTRING.format("batch_size, sequence_length"))
@add_code_sample_docstrings(
tokenizer_class=_TOKENIZER_FOR_DOC,
checkpoint="allenai/longformer-base-4096",
......@@ -1736,7 +1736,7 @@ class TFLongformerForQuestionAnswering(TFLongformerPreTrainedModel, TFQuestionAn
name="qa_outputs",
)
@add_start_docstrings_to_callable(LONGFORMER_INPUTS_DOCSTRING.format("batch_size, sequence_length"))
@add_start_docstrings_to_model_forward(LONGFORMER_INPUTS_DOCSTRING.format("batch_size, sequence_length"))
@add_code_sample_docstrings(
tokenizer_class=_TOKENIZER_FOR_DOC,
checkpoint="allenai/longformer-large-4096-finetuned-triviaqa",
......
src/transformers/modeling_tf_lxmert.py
View file @ 378142af
......@@ -28,7 +28,7 @@ from .file_utils import (
ModelOutput,
add_code_sample_docstrings,
add_start_docstrings,
add_start_docstrings_to_callable,
add_start_docstrings_to_model_forward,
replace_return_docstrings,
)
from .modeling_tf_utils import TFPreTrainedModel, get_initializer, keras_serializable, shape_list
......@@ -970,7 +970,7 @@ class TFLxmertModel(TFLxmertPreTrainedModel):
super().__init__(config, *inputs, **kwargs)
self.lxmert = TFLxmertMainLayer(config, name="lxmert")
@add_start_docstrings_to_callable(LXMERT_INPUTS_DOCSTRING)
@add_start_docstrings_to_model_forward(LXMERT_INPUTS_DOCSTRING)
@add_code_sample_docstrings(
tokenizer_class=_TOKENIZER_FOR_DOC,
checkpoint="unc-nlp/lxmert-base-uncased",
......@@ -1224,7 +1224,7 @@ class TFLxmertForPreTraining(TFLxmertPreTrainedModel):
**({"obj_labels": obj_labels} if self.config.task_obj_predict else {}),
}
@add_start_docstrings_to_callable(LXMERT_INPUTS_DOCSTRING)
@add_start_docstrings_to_model_forward(LXMERT_INPUTS_DOCSTRING)
@replace_return_docstrings(output_type=TFLxmertForPreTrainingOutput, config_class=_CONFIG_FOR_DOC)
def call(
self,
......
src/transformers/modeling_tf_mobilebert.py
View file @ 378142af
......@@ -28,7 +28,7 @@ from .file_utils import (
ModelOutput,
add_code_sample_docstrings,
add_start_docstrings,
add_start_docstrings_to_callable,
add_start_docstrings_to_model_forward,
replace_return_docstrings,
)
from .modeling_tf_outputs import (
......@@ -960,7 +960,7 @@ class TFMobileBertModel(TFMobileBertPreTrainedModel):
super().__init__(config, *inputs, **kwargs)
self.mobilebert = TFMobileBertMainLayer(config, name="mobilebert")
@add_start_docstrings_to_callable(MOBILEBERT_INPUTS_DOCSTRING.format("batch_size, sequence_length"))
@add_start_docstrings_to_model_forward(MOBILEBERT_INPUTS_DOCSTRING.format("batch_size, sequence_length"))
@add_code_sample_docstrings(
tokenizer_class=_TOKENIZER_FOR_DOC,
checkpoint="google/mobilebert-uncased",
......@@ -989,7 +989,7 @@ class TFMobileBertForPreTraining(TFMobileBertPreTrainedModel):
def get_output_embeddings(self):
return self.mobilebert.embeddings
@add_start_docstrings_to_callable(MOBILEBERT_INPUTS_DOCSTRING.format("batch_size, sequence_length"))
@add_start_docstrings_to_model_forward(MOBILEBERT_INPUTS_DOCSTRING.format("batch_size, sequence_length"))
@replace_return_docstrings(output_type=TFMobileBertForPreTrainingOutput, config_class=_CONFIG_FOR_DOC)
def call(self, inputs, **kwargs):
r"""
......@@ -1040,7 +1040,7 @@ class TFMobileBertForMaskedLM(TFMobileBertPreTrainedModel, TFMaskedLanguageModel
def get_output_embeddings(self):
return self.mobilebert.embeddings
@add_start_docstrings_to_callable(MOBILEBERT_INPUTS_DOCSTRING.format("batch_size, sequence_length"))
@add_start_docstrings_to_model_forward(MOBILEBERT_INPUTS_DOCSTRING.format("batch_size, sequence_length"))
@add_code_sample_docstrings(
tokenizer_class=_TOKENIZER_FOR_DOC,
checkpoint="google/mobilebert-uncased",
......@@ -1126,7 +1126,7 @@ class TFMobileBertForNextSentencePrediction(TFMobileBertPreTrainedModel):
self.mobilebert = TFMobileBertMainLayer(config, name="mobilebert")
self.cls = TFMobileBertOnlyNSPHead(config, name="seq_relationship___cls")
@add_start_docstrings_to_callable(MOBILEBERT_INPUTS_DOCSTRING.format("batch_size, sequence_length"))
@add_start_docstrings_to_model_forward(MOBILEBERT_INPUTS_DOCSTRING.format("batch_size, sequence_length"))
@replace_return_docstrings(output_type=TFNextSentencePredictorOutput, config_class=_CONFIG_FOR_DOC)
def call(self, inputs, **kwargs):
r"""
......@@ -1181,7 +1181,7 @@ class TFMobileBertForSequenceClassification(TFMobileBertPreTrainedModel, TFSeque
config.num_labels, kernel_initializer=get_initializer(config.initializer_range), name="classifier"
)
@add_start_docstrings_to_callable(MOBILEBERT_INPUTS_DOCSTRING.format("batch_size, sequence_length"))
@add_start_docstrings_to_model_forward(MOBILEBERT_INPUTS_DOCSTRING.format("batch_size, sequence_length"))
@add_code_sample_docstrings(
tokenizer_class=_TOKENIZER_FOR_DOC,
checkpoint="google/mobilebert-uncased",
......@@ -1268,7 +1268,7 @@ class TFMobileBertForQuestionAnswering(TFMobileBertPreTrainedModel, TFQuestionAn
config.num_labels, kernel_initializer=get_initializer(config.initializer_range), name="qa_outputs"
)
@add_start_docstrings_to_callable(MOBILEBERT_INPUTS_DOCSTRING.format("batch_size, sequence_length"))
@add_start_docstrings_to_model_forward(MOBILEBERT_INPUTS_DOCSTRING.format("batch_size, sequence_length"))
@add_code_sample_docstrings(
tokenizer_class=_TOKENIZER_FOR_DOC,
checkpoint="google/mobilebert-uncased",
......@@ -1376,7 +1376,9 @@ class TFMobileBertForMultipleChoice(TFMobileBertPreTrainedModel, TFMultipleChoic
"""
return {"input_ids": tf.constant(MULTIPLE_CHOICE_DUMMY_INPUTS)}
@add_start_docstrings_to_callable(MOBILEBERT_INPUTS_DOCSTRING.format("batch_size, num_choices, sequence_length"))
@add_start_docstrings_to_model_forward(
MOBILEBERT_INPUTS_DOCSTRING.format("batch_size, num_choices, sequence_length")
)
@add_code_sample_docstrings(
tokenizer_class=_TOKENIZER_FOR_DOC,
checkpoint="google/mobilebert-uncased",
......@@ -1499,7 +1501,7 @@ class TFMobileBertForTokenClassification(TFMobileBertPreTrainedModel, TFTokenCla
config.num_labels, kernel_initializer=get_initializer(config.initializer_range), name="classifier"
)
@add_start_docstrings_to_callable(MOBILEBERT_INPUTS_DOCSTRING.format("batch_size, sequence_length"))
@add_start_docstrings_to_model_forward(MOBILEBERT_INPUTS_DOCSTRING.format("batch_size, sequence_length"))
@add_code_sample_docstrings(
tokenizer_class=_TOKENIZER_FOR_DOC,
checkpoint="google/mobilebert-uncased",
......
src/transformers/modeling_tf_openai.py
View file @ 378142af
......@@ -27,7 +27,7 @@ from .file_utils import (
ModelOutput,
add_code_sample_docstrings,
add_start_docstrings,
add_start_docstrings_to_callable,
add_start_docstrings_to_model_forward,
replace_return_docstrings,
)
from .modeling_tf_outputs import TFBaseModelOutput, TFCausalLMOutput
......@@ -495,7 +495,7 @@ class TFOpenAIGPTModel(TFOpenAIGPTPreTrainedModel):
super().__init__(config, *inputs, **kwargs)
self.transformer = TFOpenAIGPTMainLayer(config, name="transformer")
@add_start_docstrings_to_callable(OPENAI_GPT_INPUTS_DOCSTRING)
@add_start_docstrings_to_model_forward(OPENAI_GPT_INPUTS_DOCSTRING)
@add_code_sample_docstrings(
tokenizer_class=_TOKENIZER_FOR_DOC,
checkpoint="openai-gpt",
......@@ -522,7 +522,7 @@ class TFOpenAIGPTLMHeadModel(TFOpenAIGPTPreTrainedModel, TFCausalLanguageModelin
def get_output_embeddings(self):
return self.transformer.tokens_embed
@add_start_docstrings_to_callable(OPENAI_GPT_INPUTS_DOCSTRING)
@add_start_docstrings_to_model_forward(OPENAI_GPT_INPUTS_DOCSTRING)
@add_code_sample_docstrings(
tokenizer_class=_TOKENIZER_FOR_DOC,
checkpoint="openai-gpt",
......@@ -612,7 +612,7 @@ class TFOpenAIGPTDoubleHeadsModel(TFOpenAIGPTPreTrainedModel):
def get_output_embeddings(self):
return self.transformer.tokens_embed
@add_start_docstrings_to_callable(OPENAI_GPT_INPUTS_DOCSTRING)
@add_start_docstrings_to_model_forward(OPENAI_GPT_INPUTS_DOCSTRING)
@replace_return_docstrings(output_type=TFOpenAIGPTDoubleHeadsModelOutput, config_class=_CONFIG_FOR_DOC)
def call(
self,
......
src/transformers/modeling_tf_roberta.py
View file @ 378142af
......@@ -24,7 +24,7 @@ from .file_utils import (
MULTIPLE_CHOICE_DUMMY_INPUTS,
add_code_sample_docstrings,
add_start_docstrings,
add_start_docstrings_to_callable,
add_start_docstrings_to_model_forward,
)
from .modeling_tf_outputs import (
TFBaseModelOutput,
......@@ -717,7 +717,7 @@ class TFRobertaModel(TFRobertaPreTrainedModel):
super().__init__(config, *inputs, **kwargs)
self.roberta = TFRobertaMainLayer(config, name="roberta")
@add_start_docstrings_to_callable(ROBERTA_INPUTS_DOCSTRING.format("batch_size, sequence_length"))
@add_start_docstrings_to_model_forward(ROBERTA_INPUTS_DOCSTRING.format("batch_size, sequence_length"))
@add_code_sample_docstrings(
tokenizer_class=_TOKENIZER_FOR_DOC,
checkpoint="roberta-base",
......@@ -776,7 +776,7 @@ class TFRobertaForMaskedLM(TFRobertaPreTrainedModel, TFMaskedLanguageModelingLos
def get_output_embeddings(self):
return self.lm_head.decoder
@add_start_docstrings_to_callable(ROBERTA_INPUTS_DOCSTRING.format("batch_size, sequence_length"))
@add_start_docstrings_to_model_forward(ROBERTA_INPUTS_DOCSTRING.format("batch_size, sequence_length"))
@add_code_sample_docstrings(
tokenizer_class=_TOKENIZER_FOR_DOC,
checkpoint="roberta-base",
......@@ -886,7 +886,7 @@ class TFRobertaForSequenceClassification(TFRobertaPreTrainedModel, TFSequenceCla
self.roberta = TFRobertaMainLayer(config, name="roberta")
self.classifier = TFRobertaClassificationHead(config, name="classifier")
@add_start_docstrings_to_callable(ROBERTA_INPUTS_DOCSTRING.format("batch_size, sequence_length"))
@add_start_docstrings_to_model_forward(ROBERTA_INPUTS_DOCSTRING.format("batch_size, sequence_length"))
@add_code_sample_docstrings(
tokenizer_class=_TOKENIZER_FOR_DOC,
checkpoint="roberta-base",
......@@ -978,7 +978,7 @@ class TFRobertaForMultipleChoice(TFRobertaPreTrainedModel, TFMultipleChoiceLoss)
"""
return {"input_ids": tf.constant(MULTIPLE_CHOICE_DUMMY_INPUTS)}
@add_start_docstrings_to_callable(ROBERTA_INPUTS_DOCSTRING.format("batch_size, num_choices, sequence_length"))
@add_start_docstrings_to_model_forward(ROBERTA_INPUTS_DOCSTRING.format("batch_size, num_choices, sequence_length"))
@add_code_sample_docstrings(
tokenizer_class=_TOKENIZER_FOR_DOC,
checkpoint="roberta-base",
......@@ -1096,7 +1096,7 @@ class TFRobertaForTokenClassification(TFRobertaPreTrainedModel, TFTokenClassific
config.num_labels, kernel_initializer=get_initializer(config.initializer_range), name="classifier"
)
@add_start_docstrings_to_callable(ROBERTA_INPUTS_DOCSTRING.format("batch_size, sequence_length"))
@add_start_docstrings_to_model_forward(ROBERTA_INPUTS_DOCSTRING.format("batch_size, sequence_length"))
@add_code_sample_docstrings(
tokenizer_class=_TOKENIZER_FOR_DOC,
checkpoint="roberta-base",
......@@ -1182,7 +1182,7 @@ class TFRobertaForQuestionAnswering(TFRobertaPreTrainedModel, TFQuestionAnswerin
config.num_labels, kernel_initializer=get_initializer(config.initializer_range), name="qa_outputs"
)
@add_start_docstrings_to_callable(ROBERTA_INPUTS_DOCSTRING.format("batch_size, sequence_length"))
@add_start_docstrings_to_model_forward(ROBERTA_INPUTS_DOCSTRING.format("batch_size, sequence_length"))
@add_code_sample_docstrings(
tokenizer_class=_TOKENIZER_FOR_DOC,
checkpoint="roberta-base",
......
src/transformers/modeling_tf_t5.py
View file @ 378142af
......@@ -31,7 +31,7 @@ from .file_utils import (
DUMMY_INPUTS,
DUMMY_MASK,
add_start_docstrings,
add_start_docstrings_to_callable,
add_start_docstrings_to_model_forward,
replace_return_docstrings,
)
from .modeling_tf_outputs import TFSeq2SeqLMOutput, TFSeq2SeqModelOutput
......@@ -980,7 +980,7 @@ class TFT5Model(TFT5PreTrainedModel):
def get_decoder(self):
return self.decoder
@add_start_docstrings_to_callable(T5_INPUTS_DOCSTRING)
@add_start_docstrings_to_model_forward(T5_INPUTS_DOCSTRING)
@replace_return_docstrings(output_type=TFSeq2SeqModelOutput, config_class=_CONFIG_FOR_DOC)
def call(
self,
......@@ -1177,7 +1177,7 @@ class TFT5ForConditionalGeneration(TFT5PreTrainedModel, TFCausalLanguageModeling
def get_decoder(self):
return self.decoder
@add_start_docstrings_to_callable(T5_INPUTS_DOCSTRING)
@add_start_docstrings_to_model_forward(T5_INPUTS_DOCSTRING)
@replace_return_docstrings(output_type=TFSeq2SeqLMOutput, config_class=_CONFIG_FOR_DOC)
def call(
self,
......
src/transformers/modeling_tf_transfo_xl.py
View file @ 378142af
......@@ -23,7 +23,12 @@ from typing import List, Optional, Tuple
import tensorflow as tf
from .configuration_transfo_xl import TransfoXLConfig
from .file_utils import ModelOutput, add_code_sample_docstrings, add_start_docstrings, add_start_docstrings_to_callable
from .file_utils import (
ModelOutput,
add_code_sample_docstrings,
add_start_docstrings,
add_start_docstrings_to_model_forward,
)
from .modeling_tf_transfo_xl_utilities import TFAdaptiveSoftmaxMask
from .modeling_tf_utils import TFPreTrainedModel, get_initializer, keras_serializable, shape_list
from .tokenization_utils import BatchEncoding
......@@ -803,7 +808,7 @@ class TFTransfoXLModel(TFTransfoXLPreTrainedModel):
super().__init__(config, *inputs, **kwargs)
self.transformer = TFTransfoXLMainLayer(config, name="transformer")
@add_start_docstrings_to_callable(TRANSFO_XL_INPUTS_DOCSTRING)
@add_start_docstrings_to_model_forward(TRANSFO_XL_INPUTS_DOCSTRING)
@add_code_sample_docstrings(
tokenizer_class=_TOKENIZER_FOR_DOC,
checkpoint="transfo-xl-wt103",
......@@ -873,7 +878,7 @@ class TFTransfoXLLMHeadModel(TFTransfoXLPreTrainedModel):
def init_mems(self, bsz):
return self.transformer.init_mems(bsz)
@add_start_docstrings_to_callable(TRANSFO_XL_INPUTS_DOCSTRING)
@add_start_docstrings_to_model_forward(TRANSFO_XL_INPUTS_DOCSTRING)
@add_code_sample_docstrings(
tokenizer_class=_TOKENIZER_FOR_DOC,
checkpoint="transfo-xl-wt103",
......
src/transformers/modeling_tf_xlm.py
View file @ 378142af
......@@ -32,7 +32,7 @@ from .file_utils import (
ModelOutput,
add_code_sample_docstrings,
add_start_docstrings,
add_start_docstrings_to_callable,
add_start_docstrings_to_model_forward,
)
from .modeling_tf_outputs import (
TFBaseModelOutput,
......@@ -696,7 +696,7 @@ class TFXLMModel(TFXLMPreTrainedModel):
super().__init__(config, *inputs, **kwargs)
self.transformer = TFXLMMainLayer(config, name="transformer")
@add_start_docstrings_to_callable(XLM_INPUTS_DOCSTRING.format("batch_size, sequence_length"))
@add_start_docstrings_to_model_forward(XLM_INPUTS_DOCSTRING.format("batch_size, sequence_length"))
@add_code_sample_docstrings(
tokenizer_class=_TOKENIZER_FOR_DOC,
checkpoint="xlm-mlm-en-2048",
......@@ -775,7 +775,7 @@ class TFXLMWithLMHeadModel(TFXLMPreTrainedModel):
langs = None
return {"inputs": inputs, "langs": langs}
@add_start_docstrings_to_callable(XLM_INPUTS_DOCSTRING.format("batch_size, sequence_length"))
@add_start_docstrings_to_model_forward(XLM_INPUTS_DOCSTRING.format("batch_size, sequence_length"))
@add_code_sample_docstrings(
tokenizer_class=_TOKENIZER_FOR_DOC,
checkpoint="xlm-mlm-en-2048",
......@@ -813,7 +813,7 @@ class TFXLMForSequenceClassification(TFXLMPreTrainedModel, TFSequenceClassificat
self.transformer = TFXLMMainLayer(config, name="transformer")
self.sequence_summary = TFSequenceSummary(config, initializer_range=config.init_std, name="sequence_summary")
@add_start_docstrings_to_callable(XLM_INPUTS_DOCSTRING.format("batch_size, sequence_length"))
@add_start_docstrings_to_model_forward(XLM_INPUTS_DOCSTRING.format("batch_size, sequence_length"))
@add_code_sample_docstrings(
tokenizer_class=_TOKENIZER_FOR_DOC,
checkpoint="xlm-mlm-en-2048",
......@@ -914,7 +914,7 @@ class TFXLMForMultipleChoice(TFXLMPreTrainedModel, TFMultipleChoiceLoss):
"langs": tf.constant(MULTIPLE_CHOICE_DUMMY_INPUTS),
}
@add_start_docstrings_to_callable(XLM_INPUTS_DOCSTRING.format("batch_size, num_choices, sequence_length"))
@add_start_docstrings_to_model_forward(XLM_INPUTS_DOCSTRING.format("batch_size, num_choices, sequence_length"))
@add_code_sample_docstrings(
tokenizer_class=_TOKENIZER_FOR_DOC,
checkpoint="xlm-mlm-en-2048",
......@@ -1056,7 +1056,7 @@ class TFXLMForTokenClassification(TFXLMPreTrainedModel, TFTokenClassificationLos
config.num_labels, kernel_initializer=get_initializer(config.init_std), name="classifier"
)
@add_start_docstrings_to_callable(XLM_INPUTS_DOCSTRING.format("batch_size, sequence_length"))
@add_start_docstrings_to_model_forward(XLM_INPUTS_DOCSTRING.format("batch_size, sequence_length"))
@add_code_sample_docstrings(
tokenizer_class=_TOKENIZER_FOR_DOC,
checkpoint="xlm-mlm-en-2048",
......@@ -1143,7 +1143,7 @@ class TFXLMForQuestionAnsweringSimple(TFXLMPreTrainedModel, TFQuestionAnsweringL
config.num_labels, kernel_initializer=get_initializer(config.init_std), name="qa_outputs"
)
@add_start_docstrings_to_callable(XLM_INPUTS_DOCSTRING.format("batch_size, sequence_length"))
@add_start_docstrings_to_model_forward(XLM_INPUTS_DOCSTRING.format("batch_size, sequence_length"))
@add_code_sample_docstrings(
tokenizer_class=_TOKENIZER_FOR_DOC,
checkpoint="xlm-mlm-en-2048",
......
src/transformers/modeling_tf_xlnet.py
View file @ 378142af
......@@ -30,7 +30,7 @@ from .file_utils import (
ModelOutput,
add_code_sample_docstrings,
add_start_docstrings,
add_start_docstrings_to_callable,
add_start_docstrings_to_model_forward,
replace_return_docstrings,
)
from .modeling_tf_utils import (
......@@ -1130,7 +1130,7 @@ class TFXLNetModel(TFXLNetPreTrainedModel):
super().__init__(config, *inputs, **kwargs)
self.transformer = TFXLNetMainLayer(config, name="transformer")
@add_start_docstrings_to_callable(XLNET_INPUTS_DOCSTRING.format("batch_size, sequence_length"))
@add_start_docstrings_to_model_forward(XLNET_INPUTS_DOCSTRING.format("batch_size, sequence_length"))
@add_code_sample_docstrings(
tokenizer_class=_TOKENIZER_FOR_DOC,
checkpoint="xlnet-base-cased",
......@@ -1197,7 +1197,7 @@ class TFXLNetLMHeadModel(TFXLNetPreTrainedModel, TFCausalLanguageModelingLoss):
return inputs
@add_start_docstrings_to_callable(XLNET_INPUTS_DOCSTRING.format("batch_size, sequence_length"))
@add_start_docstrings_to_model_forward(XLNET_INPUTS_DOCSTRING.format("batch_size, sequence_length"))
@replace_return_docstrings(output_type=TFXLNetLMHeadModelOutput, config_class=_CONFIG_FOR_DOC)
def call(
self,
......@@ -1314,7 +1314,7 @@ class TFXLNetForSequenceClassification(TFXLNetPreTrainedModel, TFSequenceClassif
config.num_labels, kernel_initializer=get_initializer(config.initializer_range), name="logits_proj"
)
@add_start_docstrings_to_callable(XLNET_INPUTS_DOCSTRING.format("batch_size, sequence_length"))
@add_start_docstrings_to_model_forward(XLNET_INPUTS_DOCSTRING.format("batch_size, sequence_length"))
@add_code_sample_docstrings(
tokenizer_class=_TOKENIZER_FOR_DOC,
checkpoint="xlnet-base-cased",
......@@ -1417,7 +1417,7 @@ class TFXLNetForMultipleChoice(TFXLNetPreTrainedModel, TFMultipleChoiceLoss):
"""
return {"input_ids": tf.constant(MULTIPLE_CHOICE_DUMMY_INPUTS)}
@add_start_docstrings_to_callable(XLNET_INPUTS_DOCSTRING.format("batch_size, num_choices, sequence_length"))
@add_start_docstrings_to_model_forward(XLNET_INPUTS_DOCSTRING.format("batch_size, num_choices, sequence_length"))
@add_code_sample_docstrings(
tokenizer_class=_TOKENIZER_FOR_DOC,
checkpoint="xlnet-base-cased",
......@@ -1552,7 +1552,7 @@ class TFXLNetForTokenClassification(TFXLNetPreTrainedModel, TFTokenClassificatio
config.num_labels, kernel_initializer=get_initializer(config.initializer_range), name="classifier"
)
@add_start_docstrings_to_callable(XLNET_INPUTS_DOCSTRING.format("batch_size, sequence_length"))
@add_start_docstrings_to_model_forward(XLNET_INPUTS_DOCSTRING.format("batch_size, sequence_length"))
@add_code_sample_docstrings(
tokenizer_class=_TOKENIZER_FOR_DOC,
checkpoint="xlnet-base-cased",
......@@ -1639,7 +1639,7 @@ class TFXLNetForQuestionAnsweringSimple(TFXLNetPreTrainedModel, TFQuestionAnswer
config.num_labels, kernel_initializer=get_initializer(config.initializer_range), name="qa_outputs"
)
@add_start_docstrings_to_callable(XLNET_INPUTS_DOCSTRING.format("batch_size, sequence_length"))
@add_start_docstrings_to_model_forward(XLNET_INPUTS_DOCSTRING.format("batch_size, sequence_length"))
@add_code_sample_docstrings(
tokenizer_class=_TOKENIZER_FOR_DOC,
checkpoint="xlnet-base-cased",
......
src/transformers/modeling_transfo_xl.py
View file @ 378142af
......@@ -26,7 +26,12 @@ import torch.nn as nn
import torch.nn.functional as F
from .configuration_transfo_xl import TransfoXLConfig
from .file_utils import ModelOutput, add_code_sample_docstrings, add_start_docstrings, add_start_docstrings_to_callable
from .file_utils import (
ModelOutput,
add_code_sample_docstrings,
add_start_docstrings,
add_start_docstrings_to_model_forward,
)
from .modeling_transfo_xl_utilities import ProjectedAdaptiveLogSoftmax
from .modeling_utils import PreTrainedModel
from .utils import logging
......@@ -830,7 +835,7 @@ class TransfoXLModel(TransfoXLPreTrainedModel):
return new_mems
@add_start_docstrings_to_callable(TRANSFO_XL_INPUTS_DOCSTRING)
@add_start_docstrings_to_model_forward(TRANSFO_XL_INPUTS_DOCSTRING)
@add_code_sample_docstrings(
tokenizer_class=_TOKENIZER_FOR_DOC,
checkpoint="transfo-xl-wt103",
......@@ -1018,7 +1023,7 @@ class TransfoXLLMHeadModel(TransfoXLPreTrainedModel):
def init_mems(self, bsz):
return self.transformer.init_mems(bsz)
@add_start_docstrings_to_callable(TRANSFO_XL_INPUTS_DOCSTRING)
@add_start_docstrings_to_model_forward(TRANSFO_XL_INPUTS_DOCSTRING)
@add_code_sample_docstrings(
tokenizer_class=_TOKENIZER_FOR_DOC,
checkpoint="transfo-xl-wt103",
......
src/transformers/modeling_xlm.py
View file @ 378142af
......@@ -35,7 +35,7 @@ from .file_utils import (
ModelOutput,
add_code_sample_docstrings,
add_start_docstrings,
add_start_docstrings_to_callable,
add_start_docstrings_to_model_forward,
replace_return_docstrings,
)
from .modeling_outputs import (
......@@ -486,7 +486,7 @@ class XLMModel(XLMPreTrainedModel):
for layer, heads in heads_to_prune.items():
self.attentions[layer].prune_heads(heads)
@add_start_docstrings_to_callable(XLM_INPUTS_DOCSTRING.format("batch_size, sequence_length"))
@add_start_docstrings_to_model_forward(XLM_INPUTS_DOCSTRING.format("batch_size, sequence_length"))
@add_code_sample_docstrings(
tokenizer_class=_TOKENIZER_FOR_DOC,
checkpoint="xlm-mlm-en-2048",
......@@ -703,7 +703,7 @@ class XLMWithLMHeadModel(XLMPreTrainedModel):
langs = None
return {"input_ids": input_ids, "langs": langs}
@add_start_docstrings_to_callable(XLM_INPUTS_DOCSTRING.format("batch_size, sequence_length"))
@add_start_docstrings_to_model_forward(XLM_INPUTS_DOCSTRING.format("batch_size, sequence_length"))
@add_code_sample_docstrings(
tokenizer_class=_TOKENIZER_FOR_DOC,
checkpoint="xlm-mlm-en-2048",
......@@ -781,7 +781,7 @@ class XLMForSequenceClassification(XLMPreTrainedModel):
self.init_weights()
@add_start_docstrings_to_callable(XLM_INPUTS_DOCSTRING.format("batch_size, sequence_length"))
@add_start_docstrings_to_model_forward(XLM_INPUTS_DOCSTRING.format("batch_size, sequence_length"))
@add_code_sample_docstrings(
tokenizer_class=_TOKENIZER_FOR_DOC,
checkpoint="xlm-mlm-en-2048",
......@@ -868,7 +868,7 @@ class XLMForQuestionAnsweringSimple(XLMPreTrainedModel):
self.init_weights()
@add_start_docstrings_to_callable(XLM_INPUTS_DOCSTRING.format("batch_size, sequence_length"))
@add_start_docstrings_to_model_forward(XLM_INPUTS_DOCSTRING.format("batch_size, sequence_length"))
@add_code_sample_docstrings(
tokenizer_class=_TOKENIZER_FOR_DOC,
checkpoint="xlm-mlm-en-2048",
......@@ -972,7 +972,7 @@ class XLMForQuestionAnswering(XLMPreTrainedModel):
self.init_weights()
@add_start_docstrings_to_callable(XLM_INPUTS_DOCSTRING.format("batch_size, sequence_length"))
@add_start_docstrings_to_model_forward(XLM_INPUTS_DOCSTRING.format("batch_size, sequence_length"))
@replace_return_docstrings(output_type=XLMForQuestionAnsweringOutput, config_class=_CONFIG_FOR_DOC)
def forward(
self,
......@@ -1091,7 +1091,7 @@ class XLMForTokenClassification(XLMPreTrainedModel):
self.init_weights()
@add_start_docstrings_to_callable(XLM_INPUTS_DOCSTRING.format("batch_size, sequence_length"))
@add_start_docstrings_to_model_forward(XLM_INPUTS_DOCSTRING.format("batch_size, sequence_length"))
@add_code_sample_docstrings(
tokenizer_class=_TOKENIZER_FOR_DOC,
checkpoint="xlm-mlm-en-2048",
......@@ -1184,7 +1184,7 @@ class XLMForMultipleChoice(XLMPreTrainedModel):
self.init_weights()
@add_start_docstrings_to_callable(XLM_INPUTS_DOCSTRING.format("batch_size, num_choicec, sequence_length"))
@add_start_docstrings_to_model_forward(XLM_INPUTS_DOCSTRING.format("batch_size, num_choicec, sequence_length"))
@add_code_sample_docstrings(
tokenizer_class=_TOKENIZER_FOR_DOC,
checkpoint="xlm-mlm-en-2048",
......
src/transformers/modeling_xlnet.py
View file @ 378142af
......@@ -32,7 +32,7 @@ from .file_utils import (
ModelOutput,
add_code_sample_docstrings,
add_start_docstrings,
add_start_docstrings_to_callable,
add_start_docstrings_to_model_forward,
replace_return_docstrings,
)
from .modeling_utils import (
......@@ -1064,7 +1064,7 @@ class XLNetModel(XLNetPreTrainedModel):
pos_emb = pos_emb.to(self.device)
return pos_emb
@add_start_docstrings_to_callable(XLNET_INPUTS_DOCSTRING.format("batch_size, sequence_length"))
@add_start_docstrings_to_model_forward(XLNET_INPUTS_DOCSTRING.format("batch_size, sequence_length"))
@add_code_sample_docstrings(
tokenizer_class=_TOKENIZER_FOR_DOC,
checkpoint="xlnet-base-cased",
......@@ -1342,7 +1342,7 @@ class XLNetLMHeadModel(XLNetPreTrainedModel):
return inputs
@add_start_docstrings_to_callable(XLNET_INPUTS_DOCSTRING.format("batch_size, sequence_length"))
@add_start_docstrings_to_model_forward(XLNET_INPUTS_DOCSTRING.format("batch_size, sequence_length"))
@replace_return_docstrings(output_type=XLNetLMHeadModelOutput, config_class=_CONFIG_FOR_DOC)
def forward(
self,
......@@ -1465,7 +1465,7 @@ class XLNetForSequenceClassification(XLNetPreTrainedModel):
self.init_weights()
@add_start_docstrings_to_callable(XLNET_INPUTS_DOCSTRING.format("batch_size, sequence_length"))
@add_start_docstrings_to_model_forward(XLNET_INPUTS_DOCSTRING.format("batch_size, sequence_length"))
@add_code_sample_docstrings(
tokenizer_class=_TOKENIZER_FOR_DOC,
checkpoint="xlnet-base-cased",
......@@ -1558,7 +1558,7 @@ class XLNetForTokenClassification(XLNetPreTrainedModel):
self.init_weights()
@add_start_docstrings_to_callable(XLNET_INPUTS_DOCSTRING.format("batch_size, sequence_length"))
@add_start_docstrings_to_model_forward(XLNET_INPUTS_DOCSTRING.format("batch_size, sequence_length"))
@add_code_sample_docstrings(
tokenizer_class=_TOKENIZER_FOR_DOC,
checkpoint="xlnet-base-cased",
......@@ -1655,7 +1655,7 @@ class XLNetForMultipleChoice(XLNetPreTrainedModel):
self.init_weights()
@add_start_docstrings_to_callable(XLNET_INPUTS_DOCSTRING.format("batch_size, num_choices, sequence_length"))
@add_start_docstrings_to_model_forward(XLNET_INPUTS_DOCSTRING.format("batch_size, num_choices, sequence_length"))
@add_code_sample_docstrings(
tokenizer_class=_TOKENIZER_FOR_DOC,
checkpoint="xlnet-base-cased",
......@@ -1756,7 +1756,7 @@ class XLNetForQuestionAnsweringSimple(XLNetPreTrainedModel):
self.init_weights()
@add_start_docstrings_to_callable(XLNET_INPUTS_DOCSTRING.format("batch_size, sequence_length"))
@add_start_docstrings_to_model_forward(XLNET_INPUTS_DOCSTRING.format("batch_size, sequence_length"))
@add_code_sample_docstrings(
tokenizer_class=_TOKENIZER_FOR_DOC,
checkpoint="xlnet-base-cased",
......@@ -1868,7 +1868,7 @@ class XLNetForQuestionAnswering(XLNetPreTrainedModel):
self.init_weights()
@add_start_docstrings_to_callable(XLNET_INPUTS_DOCSTRING.format("batch_size, sequence_length"))
@add_start_docstrings_to_model_forward(XLNET_INPUTS_DOCSTRING.format("batch_size, sequence_length"))
@replace_return_docstrings(output_type=XLNetForQuestionAnsweringOutput, config_class=_CONFIG_FOR_DOC)
def forward(
self,
......
templates/adding_a_new_model/modeling_tf_xxx.py
View file @ 378142af
......@@ -26,7 +26,7 @@ from .file_utils import (
MULTIPLE_CHOICE_DUMMY_INPUTS,
add_code_sample_docstrings,
add_start_docstrings,
add_start_docstrings_to_callable,
add_start_docstrings_to_model_forward,
)
from .modeling_tf_outputs import (
TFBaseModelOutputWithPooling,
......@@ -360,7 +360,7 @@ class TFXxxModel(TFXxxPreTrainedModel):
super().__init__(config, *inputs, **kwargs)
self.transformer = TFXxxMainLayer(config, name="transformer")
@add_start_docstrings_to_callable(XXX_INPUTS_DOCSTRING.format("batch_size, sequence_length"))
@add_start_docstrings_to_model_forward(XXX_INPUTS_DOCSTRING.format("batch_size, sequence_length"))
@add_code_sample_docstrings(
tokenizer_class=_TOKENIZER_FOR_DOC,
checkpoint="xxx-base-cased",
......@@ -383,7 +383,7 @@ class TFXxxForMaskedLM(TFXxxPreTrainedModel, TFMaskedLanguageModelingLoss):
self.transformer = TFXxxMainLayer(config, name="transformer")
self.mlm = TFXxxMLMHead(config, self.transformer.embeddings, name="mlm")
@add_start_docstrings_to_callable(XXX_INPUTS_DOCSTRING.format("batch_size, sequence_length"))
@add_start_docstrings_to_model_forward(XXX_INPUTS_DOCSTRING.format("batch_size, sequence_length"))
@add_code_sample_docstrings(
tokenizer_class=_TOKENIZER_FOR_DOC,
checkpoint="xxx-base-cased",
......@@ -465,7 +465,7 @@ class TFXxxForSequenceClassification(TFXxxPreTrainedModel, TFSequenceClassificat
config.num_labels, kernel_initializer=get_initializer(config.initializer_range), name="classifier"
)
@add_start_docstrings_to_callable(XXX_INPUTS_DOCSTRING.format("batch_size, sequence_length"))
@add_start_docstrings_to_model_forward(XXX_INPUTS_DOCSTRING.format("batch_size, sequence_length"))
@add_code_sample_docstrings(
tokenizer_class=_TOKENIZER_FOR_DOC,
checkpoint="xxx-base-cased",
......@@ -557,7 +557,7 @@ class TFXxxForMultipleChoice(TFXxxPreTrainedModel, TFMultipleChoiceLoss):
"""
return {"input_ids": tf.constant(MULTIPLE_CHOICE_DUMMY_INPUTS)}
@add_start_docstrings_to_callable(XXX_INPUTS_DOCSTRING.format("batch_size, num_choices, sequence_length"))
@add_start_docstrings_to_model_forward(XXX_INPUTS_DOCSTRING.format("batch_size, num_choices, sequence_length"))
@add_code_sample_docstrings(
tokenizer_class=_TOKENIZER_FOR_DOC,
checkpoint="xxx-base-cased",
......@@ -680,7 +680,7 @@ class TFXxxForTokenClassification(TFXxxPreTrainedModel, TFTokenClassificationLos
config.num_labels, kernel_initializer=get_initializer(config.initializer_range), name="classifier"
)
@add_start_docstrings_to_callable(XXX_INPUTS_DOCSTRING.format("batch_size, sequence_length"))
@add_start_docstrings_to_model_forward(XXX_INPUTS_DOCSTRING.format("batch_size, sequence_length"))
@add_code_sample_docstrings(
tokenizer_class=_TOKENIZER_FOR_DOC,
checkpoint="xxx-base-cased",
......@@ -761,7 +761,7 @@ class TFXxxForQuestionAnswering(TFXxxPreTrainedModel, TFQuestionAnsweringLoss):
config.num_labels, kernel_initializer=get_initializer(config.initializer_range), name="qa_outputs"
)
@add_start_docstrings_to_callable(XXX_INPUTS_DOCSTRING.format("batch_size, sequence_length"))
@add_start_docstrings_to_model_forward(XXX_INPUTS_DOCSTRING.format("batch_size, sequence_length"))
@add_code_sample_docstrings(
tokenizer_class=_TOKENIZER_FOR_DOC,
checkpoint="xxx-base-cased",
......
templates/adding_a_new_model/modeling_xxx.py
View file @ 378142af
......@@ -26,7 +26,7 @@ from torch import nn
from torch.nn import CrossEntropyLoss, MSELoss
from .configuration_xxx import XxxConfig
from .file_utils import add_code_sample_docstrings, add_start_docstrings, add_start_docstrings_to_callable
from .file_utils import add_code_sample_docstrings, add_start_docstrings, add_start_docstrings_to_model_forward
from .modeling_outputs import (
BaseModelOutputWithPooling,
MaskedLMOutput,
......@@ -309,7 +309,7 @@ class XxxModel(XxxPreTrainedModel):
for layer, heads in heads_to_prune.items():
self.encoder.layer[layer].attention.prune_heads(heads)
@add_start_docstrings_to_callable(XXX_INPUTS_DOCSTRING.format("batch_size, sequence_length"))
@add_start_docstrings_to_model_forward(XXX_INPUTS_DOCSTRING.format("batch_size, sequence_length"))
@add_code_sample_docstrings(
tokenizer_class=_TOKENIZER_FOR_DOC,
checkpoint="xxx-base-uncased",
......@@ -391,7 +391,7 @@ class XxxForMaskedLM(XxxPreTrainedModel):
def get_output_embeddings(self):
return self.lm_head
@add_start_docstrings_to_callable(XXX_INPUTS_DOCSTRING.format("batch_size, sequence_length"))
@add_start_docstrings_to_model_forward(XXX_INPUTS_DOCSTRING.format("batch_size, sequence_length"))
@add_code_sample_docstrings(
tokenizer_class=_TOKENIZER_FOR_DOC,
checkpoint="xxx-base-uncased",
......@@ -468,7 +468,7 @@ class XxxForSequenceClassification(XxxPreTrainedModel):
self.init_weights()
@add_start_docstrings_to_callable(XXX_INPUTS_DOCSTRING.format("batch_size, sequence_length"))
@add_start_docstrings_to_model_forward(XXX_INPUTS_DOCSTRING.format("batch_size, sequence_length"))
@add_code_sample_docstrings(
tokenizer_class=_TOKENIZER_FOR_DOC,
checkpoint="xxx-base-uncased",
......@@ -551,7 +551,7 @@ class XxxForMultipleChoice(XxxPreTrainedModel):
self.init_weights()
@add_start_docstrings_to_callable(XXX_INPUTS_DOCSTRING.format("batch_size, num_choices, sequence_length"))
@add_start_docstrings_to_model_forward(XXX_INPUTS_DOCSTRING.format("batch_size, num_choices, sequence_length"))
@add_code_sample_docstrings(
tokenizer_class=_TOKENIZER_FOR_DOC,
checkpoint="xxx-base-uncased",
......@@ -641,7 +641,7 @@ class XxxForTokenClassification(XxxPreTrainedModel):
self.init_weights()
@add_start_docstrings_to_callable(XXX_INPUTS_DOCSTRING.format("batch_size, sequence_length"))
@add_start_docstrings_to_model_forward(XXX_INPUTS_DOCSTRING.format("batch_size, sequence_length"))
@add_code_sample_docstrings(
tokenizer_class=_TOKENIZER_FOR_DOC,
checkpoint="xxx-base-uncased",
......@@ -726,7 +726,7 @@ class XxxForQuestionAnswering(XxxPreTrainedModel):
self.init_weights()
@add_start_docstrings_to_callable(XXX_INPUTS_DOCSTRING.format("batch_size, sequence_length"))
@add_start_docstrings_to_model_forward(XXX_INPUTS_DOCSTRING.format("batch_size, sequence_length"))
@add_code_sample_docstrings(
tokenizer_class=_TOKENIZER_FOR_DOC,
checkpoint="xxx-base-uncased",
......
Markdown is supported
0% or .
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment