+ jXRRt^RIt^RIHtHtHt^RIHtHtH t H t H t H t H t ^RIHt^RIHtHt^RIHt^RIHtHtHtHtHtHtHtHt!RR ]4t!R R ]4t!R R ]4t !RR]4t!!RR]4t"!RR]4t#!RR]4t$!RR]4t%.ROt&R#)zPyTorch CamemBERT model.N)BCEWithLogitsLossCrossEntropyLossMSELoss),BaseModelOutputWithPoolingAndCrossAttentions!CausalLMOutputWithCrossAttentionsMaskedLMOutputMultipleChoiceModelOutputQuestionAnsweringModelOutputSequenceClassifierOutputTokenClassifierOutput)Unpack)TransformersKwargsauto_docstring)can_return_tuple)RobertaForCausalLMRobertaForMaskedLMRobertaForMultipleChoiceRobertaForQuestionAnswering RobertaForSequenceClassificationRobertaForTokenClassification RobertaModelRobertaPreTrainedModelc]tRt^,tRtRtR#)CamembertPreTrainedModelrobertaN)__name__ __module__ __qualname____firstlineno__base_model_prefix__static_attributes__rp/home/luca/screen-checking/.venv/lib/python3.14/site-packages/transformers/models/camembert/modular_camembert.pyrr,s!r"rc]tRt^0tRtR#)CamembertModelrN)rrrrr!rr"r#r%r%0sr"r%chaa]tRt^4toRRRR/tV3Rlt]]R V3RlRll44tRt Vt V;t #) CamembertForMaskedLMlm_head.decoder.weight)roberta.embeddings.word_embeddings.weightlm_head.decoder.bias lm_head.biascN<\SV`V4V=\VRR7VnR#Fadd_pooling_layerNsuper__init__ camembertr%rselfconfig __class__s&&r#r2CamembertForMaskedLM.__init__:#  N%fF r"c<V^8dQhRS[PR,RS[PR,RS[PR,RS[PR,RS[PR,RS[PR,RS[PR,R S[PR,R S[S[,R S[S[P ,S[,/ #)  input_idsNattention_masktoken_type_ids position_ids inputs_embedsencoder_hidden_statesencoder_attention_masklabelskwargsreturn)torch LongTensor FloatTensorr r tupleTensorr)format __classdict__s"r# __annotate__!CamembertForMaskedLM.__annotate__Bs5 5 ##d*5 ))D05 ((4/ 5 &&- 5 ((4/ 5  %00475 !& 1 1D 85   4'5 +,5  u|| ~ -5 r"c  VP!V3RVRVRVRVRVRVRR/V Bp V ^,p VPV 4p R p VebVPV P4p\ 4pV!V P R VP P4VP R 44p \V V V PV PR 7#) a token_type_ids (`torch.LongTensor` of shape `(batch_size, sequence_length)`, *optional*): Segment token indices to indicate first and second portions of the inputs. Indices are selected in `[0,1]`: - 0 corresponds to a *sentence A* token, - 1 corresponds to a *sentence B* token. This parameter can only be used when the model is initialized with `type_vocab_size` parameter with value >= 2. All the value in this tensor should be always < type_vocab_size. [What are token type IDs?](../glossary#token-type-ids) labels (`torch.LongTensor` of shape `(batch_size, sequence_length)`, *optional*): Labels for computing the masked language modeling loss. Indices should be in `[-100, 0, ..., config.vocab_size]` (see `input_ids` docstring) Tokens with indices set to `-100` are ignored (masked), the loss is only computed for the tokens with labels in `[0, ..., config.vocab_size]` r=r>r?r@rArB return_dictTNlosslogits hidden_states attentions) rlm_headtodevicerviewr6 vocab_sizerrTrU)r5r<r=r>r?r@rArBrCrDoutputssequence_outputprediction_scoresmasked_lm_lossloss_fcts&&&&&&&&&, r#forwardCamembertForMaskedLM.forward@s:,,   )  *  &  (  #8  $:      "!* LL9  YY0778F')H%&7&<&r?r@rCrDrE)rFrGrHr r rIrJr )rKrLs"r#rM/CamembertForSequenceClassification.__annotate__sC C ##d*C ))D0C ((4/ C &&- C ((4/ C   4'C +,C  u|| 7 7C r"c (VP!V3RVRVRVRVRR/VBpV^,p VPV 4p Rp VEeVPV P4pVPP fVP ^8XdRVPnMrVP ^8dQVP\P8Xg VP\P8XdR VPnMR VPnVPP R8XdK\4p VP ^8Xd'V !V P4VP44p MV !W4p MVPP R 8Xd>\4p V !V PR VP 4VPR 44p M-VPP R 8Xd\4p V !W4p \!V V VP"VP$R 7#) a token_type_ids (`torch.LongTensor` of shape `(batch_size, sequence_length)`, *optional*): Segment token indices to indicate first and second portions of the inputs. Indices are selected in `[0,1]`: - 0 corresponds to a *sentence A* token, - 1 corresponds to a *sentence B* token. This parameter can only be used when the model is initialized with `type_vocab_size` parameter with value >= 2. All the value in this tensor should be always < type_vocab_size. [What are token type IDs?](../glossary#token-type-ids) labels (`torch.LongTensor` of shape `(batch_size,)`, *optional*): Labels for computing the sequence classification/regression loss. Indices should be in `[0, ..., config.num_labels - 1]`. If `config.num_labels == 1` a regression loss is computed (Mean-Square loss), If `config.num_labels > 1` a classification loss is computed (Cross-Entropy). r=r>r?r@rPTN regressionsingle_label_classificationmulti_label_classificationrQrV)r classifierrXrYr6 problem_type num_labelsdtyperFlongintrsqueezerrZrr rTrU r5r<r=r>r?r@rCrDr\r]rSrRr`s &&&&&&&, r#ra*CamembertForSequenceClassification.forwards6,,  ) * &  (     "!*1  YYv}}-F{{''/??a'/;DKK,__q(fllejj.HFLL\a\e\eLe/LDKK,/KDKK,{{''<7"9??a'#FNN$4fnn6FGD#F3D))-JJ+- B @&++b/R))-II,./'!//))   r"rcNNNNNN rrrrr2rrrar!rfrgrhs@@r#rjrjzs5G C C C C r"rjc\aa]tRt^toV3Rlt]]RV3RlRll44tRtVt V;t #)CamembertForMultipleChoicecN<\SV`V4V=\VRR7VnR#)Tr.Nr0r4s&&r#r2#CamembertForMultipleChoice.__init__s#  N%fE r"cZ<V^8dQhRS[PR,RS[PR,RS[PR,RS[PR,RS[PR,RS[PR,RS[S[,R S[S[P ,S[,/#) r;r<Nr>r=rCr?r@rDrE)rFrGrHr r rIrJr)rKrLs"r#rM'CamembertForMultipleChoice.__annotate__sP P ##d*P ((4/P ))D0 P   4' P &&- P ((4/P +,P  u|| 8 8P r"c ^VeVP^,MVP^,pVe"VPR VPR 44MRp Ve"VPR VPR 44MRp Ve"VPR VPR 44MRp Ve"VPR VPR 44MRp Ve2VPR VPR 4VPR 44MRp VP!V 3RV RV RV RV RR/VBpV^,pVP V4pVP V4pVPR V4pRpVe/VP VP4p\4pV!VV4p\VVVPVPR7#) aO input_ids (`torch.LongTensor` of shape `(batch_size, num_choices, sequence_length)`): Indices of input sequence tokens in the vocabulary. Indices can be obtained using [`AutoTokenizer`]. See [`PreTrainedTokenizer.encode`] and [`PreTrainedTokenizer.__call__`] for details. [What are input IDs?](../glossary#input-ids) token_type_ids (`torch.LongTensor` of shape `(batch_size, num_choices, sequence_length)`, *optional*): Segment token indices to indicate first and second portions of the inputs. Indices are selected in `[0,1]`: - 0 corresponds to a *sentence A* token, - 1 corresponds to a *sentence B* token. This parameter can only be used when the model is initialized with `type_vocab_size` parameter with value >= 2. All the value in this tensor should be always < type_vocab_size. [What are token type IDs?](../glossary#token-type-ids) labels (`torch.LongTensor` of shape `(batch_size,)`, *optional*): Labels for computing the multiple choice classification loss. Indices should be in `[0, ..., num_choices-1]` where `num_choices` is the size of the second dimension of the input tensors. (See `input_ids` above) position_ids (`torch.LongTensor` of shape `(batch_size, num_choices, sequence_length)`, *optional*): Indices of positions of each input sequence tokens in the position embeddings. Selected in the range `[0, config.max_position_embeddings - 1]`. [What are position IDs?](../glossary#position-ids) inputs_embeds (`torch.FloatTensor` of shape `(batch_size, num_choices, sequence_length, hidden_size)`, *optional*): Optionally, instead of passing `input_ids` you can choose to directly pass an embedded representation. This is useful if you want more control over how to convert `input_ids` indices into associated vectors than the model's internal embedding lookup matrix. Nr?r>r=r@rPTrQrV) shaperZsizerdropoutrtrXrYrrrTrU)r5r<r>r=rCr?r@rD num_choicesflat_input_idsflat_position_idsflat_token_type_idsflat_attention_maskflat_inputs_embedsr\ pooled_outputrSreshaped_logitsrRr`s&&&&&&&, r#ra"CamembertForMultipleChoice.forwardsV-6,Aiooa(}GZGZ[\G] CLCXINN2,>?^bLXLdL--b,2C2CB2GHjnR`Rln11"n6I6I"6MNrvR`Rln11"n6I6I"6MNrv(   r=#5#5b#9=;M;Mb;Q R  ,,  * / /  -        ]3 / ++b+6  YY556F')HOV4D("!//))   r"rcr}r~rhs@@r#rrs5F P P P P r"rc\aa]tRtRtoV3Rlt]]RV3RlRll44tRtVt V;t #)CamembertForTokenClassificationi%cN<\SV`V4V=\VRR7VnR#r-r0r4s&&r#r2(CamembertForTokenClassification.__init__&r9r"cZ<V^8dQhRS[PR,RS[PR,RS[PR,RS[PR,RS[PR,RS[PR,RS[S[,R S[S[P ,S[,/#rn)rFrGrHr r rIrJr )rKrLs"r#rM,CamembertForTokenClassification.__annotate__.s2 2 ##d*2 ))D02 ((4/ 2 &&- 2 ((4/ 2   4'2 +,2  u|| 4 42 r"c VP!V3RVRVRVRVRR/VBpV^,p VPV 4p VPV 4p Rp VeXVPV P4p\ 4p V !V P R VP4VP R 44p \V V VPVPR7#) a token_type_ids (`torch.LongTensor` of shape `(batch_size, sequence_length)`, *optional*): Segment token indices to indicate first and second portions of the inputs. Indices are selected in `[0,1]`: - 0 corresponds to a *sentence A* token, - 1 corresponds to a *sentence B* token. This parameter can only be used when the model is initialized with `type_vocab_size` parameter with value >= 2. All the value in this tensor should be always < type_vocab_size. [What are token type IDs?](../glossary#token-type-ids) labels (`torch.LongTensor` of shape `(batch_size, sequence_length)`, *optional*): Labels for computing the token classification loss. Indices should be in `[0, ..., config.num_labels - 1]`. r=r>r?r@rPTNrQrV) rrrtrXrYrrZrvr rTrUr{s &&&&&&&, r#ra'CamembertForTokenClassification.forward,s2,,  ) * &  (     "!*,,71  YYv}}-F')HFKKDOOr?r@start_positions end_positionsrDrE)rFrGrHr r rIrJr )rKrLs"r#rM*CamembertForQuestionAnswering.__annotate__ls> > ##d*> ))D0> ((4/ > &&- > ((4/ > ))D0> ''$.> +,>  u|| ; ;> r"c VP!V3RVRVRVRVRR/VBp V ^,p VPV 4p V P^R R7wrV PR 4P 4p V PR 4P 4p RpVeVe\ VP 44^8dVPR 4p\ VP 44^8dVPR 4pV P ^4pVP^V4pVP^V4p\VR 7pV!W4pV!W4pVV,^, p\VV V V PV PR 7#) a token_type_ids (`torch.LongTensor` of shape `(batch_size, sequence_length)`, *optional*): Segment token indices to indicate first and second portions of the inputs. Indices are selected in `[0,1]`: - 0 corresponds to a *sentence A* token, - 1 corresponds to a *sentence B* token. This parameter can only be used when the model is initialized with `type_vocab_size` parameter with value >= 2. All the value in this tensor should be always < type_vocab_size. [What are token type IDs?](../glossary#token-type-ids) r=r>r?r@rPT)dimN) ignore_index)rR start_logits end_logitsrTrUrV) r qa_outputssplitrz contiguouslenrclamprr rTrU)r5r<r=r>r?r@rrrDr\r]rSrr total_loss ignored_indexr` start_lossend_losss&&&&&&&&, r#ra%CamembertForQuestionAnswering.forwardjs0,,  ) * &  (     "!*1#)<<r<#: #++B/::< ''+668   &=+D?'')*Q."1"9"9""==%%'(1, - 5 5b 9 (--a0M-33A}EO)//=AM']CH!,@J :H$x/14J+%!!//))   r"rc)NNNNNNNr~rhs@@r#rrcs1G > > > > r"rchaa]tRtRtoRRRR/tV3Rlt]]R V3RlRll44tR t Vt V;t #) CamembertForCausalLMir(r)r*r+cN<\SV`V4V=\VRR7VnR#r-r0r4s&&r#r2CamembertForCausalLM.__init__r9r"c0<V^8dQhRS[PR,RS[PR,RS[PR,RS[PR,RS[PR,RS[PR,RS[PR,R S[PR,R S[S[S[P,,R,R S[R,R S[S[P ,R S[S[,RS[S[P ,S[ ,/ #)r;r<Nr=r>r?r@rArBrCpast_key_values use_cachelogits_to_keeprDrE) rFrGrHrIboolryrJr r r)rKrLs"r#rM!CamembertForCausalLM.__annotate__s+O O ##d*O ))D0O ((4/ O &&- O ((4/ O  %0047O !& 1 1D 8O   4'O uU%6%6784?O $;O ell*O +,O  u|| @ @O r"c  VeRp VP!V3RVRVRVRVRVRVR V R V R R / V Bp V Pp\V \4'd\ V )R4MT pVP VR VR 3,4pRpVe-VP !RRVRVRVPP/V Bp\VVV PV PV PV PR7#)a token_type_ids (`torch.LongTensor` of shape `(batch_size, sequence_length)`, *optional*): Segment token indices to indicate first and second portions of the inputs. Indices are selected in `[0,1]`: - 0 corresponds to a *sentence A* token, - 1 corresponds to a *sentence B* token. This parameter can only be used when the model is initialized with `type_vocab_size` parameter with value >= 2. All the value in this tensor should be always < type_vocab_size. [What are token type IDs?](../glossary#token-type-ids) labels (`torch.LongTensor` of shape `(batch_size, sequence_length)`, *optional*): Labels for computing the left-to-right language modeling loss (next word prediction). Indices should be in `[-100, 0, ..., config.vocab_size]` (see `input_ids` docstring) Tokens with indices set to `-100` are ignored (masked), the loss is only computed for the tokens with labels in `[0, ..., config.vocab_size]` Example: ```python >>> from transformers import AutoTokenizer, CamembertForCausalLM, AutoConfig >>> import torch >>> tokenizer = AutoTokenizer.from_pretrained("almanach/camembert-base") >>> config = AutoConfig.from_pretrained("almanach/camembert-base") >>> config.is_decoder = True >>> model = CamembertForCausalLM.from_pretrained("almanach/camembert-base", config=config) >>> inputs = tokenizer("Hello, my dog is cute", return_tensors="pt") >>> outputs = model(**inputs) >>> prediction_logits = outputs.logits ```NFr=r>r?r@rArBrrrPT:NNNrSrCr[)rRrSrrTrUcross_attentionsr)rlast_hidden_state isinstanceryslicerW loss_functionr6r[rrrTrUr)r5r<r=r>r?r@rArBrCrrrrDr\rT slice_indicesrSrRs&&&&&&&&&&&&, r#raCamembertForCausalLM.forwardsB`  I@D  A ) A * A & A ( A #8 A $: A , A   A  A  A  11 8B>SV8W8W~ot4]k mA}a,?@A  %%pVpFpt{{OeOepiopD0#33!//))$55   r"rc) NNNNNNNNNNrdrhs@@r#rrsH "M G O O O O r"r)rr'rrrjrr%r)'__doc__rFtorch.nnrrrmodeling_outputsrrrrr r r processing_utilsr utilsr r utils.genericrroberta.modeling_robertarrrrrrrrrr%r'rjrrrr__all__rr"r#rs AA'7-   "5" \ C -C LL )IL ^Y !9Y x; &C; |G $?G T] -] @ r"