\955r=/S?Qr>g)BzPyTorch MarkupLM model.N)CallableOptionalUnion)nn)BCEWithLogitsLossCrossEntropyLossMSELoss)ACT2FN)GradientCheckpointingLayer)BaseModelOutputBaseModelOutputWithPoolingMaskedLMOutputQuestionAnsweringModelOutputSequenceClassifierOutputTokenClassifierOutput)ALL_ATTENTION_FUNCTIONSPreTrainedModelapply_chunking_to_forward find_pruneable_heads_and_indicesprune_linear_layer)auto_docstringcan_return_tuplelogging)MarkupLMConfigc6^\rSrSrSrU4SjrSSjrSrU=r$)XPathEmbeddings1zConstruct the embeddings from xpath tags and subscripts. We drop tree-id in this version, as its info can be covered by xpath. c >[TU]5 URUl[R"UR UR-UR 5Ul[R"UR5Ul [R"5Ul [R"UR UR-SUR -5Ul [R"SUR -UR 5Ul[R"[!UR5Vs/sH.n[R""UR$UR 5PM0 sn5Ul[R"[!UR5Vs/sH.n[R""UR(UR 5PM0 sn5Ulgs snfs snf)N)super__init__ max_depthrLinearxpath_unit_hidden_size hidden_sizexpath_unitseq2_embeddingsDropouthidden_dropout_probdropoutReLU activationxpath_unitseq2_inner inner2emb ModuleListrange Embeddingmax_xpath_tag_unit_embeddingsxpath_tag_sub_embeddingsmax_xpath_subs_unit_embeddingsxpath_subs_sub_embeddings)selfconfig_ __class__s f/var/www/html/shao/venv/lib/python3.13/site-packages/transformers/models/markuplm/modeling_markuplm.pyr#XPathEmbeddings.__init__7sd )))+63P3PSWSaSa3acicucu)v&zz&"<"<= '')$&IIf.K.Kdnn.\^_bhbtbt^t$u!1v'9'9#96;M;MN(* t~~. .A VAA6C`C`a. ) %*,t~~. .A VBBFDaDab. * &   s 74G4G$c /n/n[UR5H_nURURU"USS2SS2U455 URURU"USS2SS2U455 Ma [ R "USS9n[ R "USS9nX4-nURURURURU5555nU$)Ndim) r1r$appendr4r6torchcatr/r+r-r.)r7xpath_tags_seqxpath_subs_seqxpath_tags_embeddingsxpath_subs_embeddingsixpath_embeddingss r;forwardXPathEmbeddings.forwardQs " "t~~&A ! ( ()F)Fq)I.YZ\]_`Y`Ja)b c ! ( ()G)G)J>Z[]^`aZaKb)c d'!& *?R H % *?R H0H>>$,,ttG`G`aqGr7s*tu)r-r+r/r$r6r4r(r.)NN) __name__ __module__ __qualname____firstlineno____doc__r#rJ__static_attributes__ __classcell__r:s@r;rr1s  4  rLrcURU5R5n[R"USS9R U5U-U-nUR 5U-$)z Replace non-padding symbols with their position numbers. Position numbers begin at padding_idx+1. Padding symbols are ignored. This is modified from fairseq's `utils.make_positions`. Args: x: torch.Tensor x: Returns: torch.Tensor rr?)neintrBcumsumtype_aslong) input_ids padding_idxpast_key_values_lengthmaskincremental_indicess r;"create_position_ids_from_input_idsr`dsW << $ ( ( *D <<!4<[TU]5 Xl[R"UR UR URS9Ul[R"URUR 5Ul URUl [U5Ul [R"URUR 5Ul[R "UR UR"S9Ul[R$"UR&5UlUR+S[,R."UR5R1S5SS9 URUl[R"URUR UR2S9Ul g)N)r\eps position_ids)rr>F) persistent)r"r#r8rr2 vocab_sizer' pad_token_idword_embeddingsmax_position_embeddingsposition_embeddingsr$rrItype_vocab_sizetoken_type_embeddings LayerNormlayer_norm_epsr)r*r+register_bufferrBarangeexpandr\r7r8r:s r;r#MarkupLMEmbeddings.__init__ws8  !||F,=,=v?Q?Q_e_r_rs#%<<0N0NPVPbPb#c )) / 7%'\\&2H2H&J\J\%]"f&8&8f>S>STzz&"<"<=   ELL)G)GHOOPWXej  "..#%<<  * *F,>,>DL\L\$  rLcUR5SSnUSn[R"URS-X0R-S-[RUR S9nUR S5RU5$)z We are provided embeddings directly. We cannot infer which are padded so just generate sequential position ids. Args: inputs_embeds: torch.Tensor Returns: torch.Tensor Nr>rdtypedevicer)sizerBrsr\rZrz unsqueezert)r7 inputs_embeds input_shapesequence_lengthrgs r;&create_position_ids_from_inputs_embeds9MarkupLMEmbeddings.create_position_ids_from_inputs_embedss~$((*3B/ %a.||   q /4D4D"Dq"HPUPZPZcpcwcw %%a(// <rx)r{rzr`r\rrBzerosrZrkr8 tag_pad_idonestuplelistr$ subs_pad_idrmrorIrpr+)r7r[rDrEtoken_type_idsrgr}r]r~rzwords_embeddingsrmrorI embeddingss r;rJMarkupLMEmbeddings.forwards  #..*K',,.s3K%.%:!! @T@T  $A)M]M]_uv #JJ=Y  !"[[EJJvVN   00;M  !![[33ejjd;'4>>*::;5::V\7N  !![[44uzzd;'4>>*::;5::V\8N)"66|D $ : :> J00P%;>SSVff ^^J/ \\*- rL) rpr8r+r$r\rmrorkrI)NNNNNNr) rMrNrOrPrQr#rrJrRrSrTs@r;rbrbts1Q 2=& 22rLrbcz^\rSrSrU4SjrS\R S\R S\R 4SjrSrU=r $)MarkupLMSelfOutputc(>[TU]5 [R"URUR5Ul[R "URURS9Ul[R"UR5Ul gNre) r"r#rr%r'denserprqr)r*r+rus r;r#MarkupLMSelfOutput.__init__s` YYv1163E3EF f&8&8f>S>STzz&"<"<= rL hidden_states input_tensorreturncpURU5nURU5nURX-5nU$Nrr+rpr7rrs r;rJMarkupLMSelfOutput.forward5 =1  ]3 }'CD rLrprr+ rMrNrOrPr#rBTensorrJrRrSrTs@r;rr6> U\\RWR^R^rLrcb^\rSrSrU4SjrS\R S\R 4SjrSrU=r $)MarkupLMIntermediatec>[TU]5 [R"URUR 5Ul[UR[5(a[URUl gURUl gr) r"r#rr%r'intermediate_sizer isinstance hidden_actstrr intermediate_act_fnrus r;r#MarkupLMIntermediate.__init__s` YYv1163K3KL f'' - -'-f.?.?'@D $'-'8'8D $rLrrcJURU5nURU5nU$rrrr7rs r;rJMarkupLMIntermediate.forwards& =1 00? rLrrrTs@r;rrs(9U\\ellrLrcz^\rSrSrU4SjrS\R S\R S\R 4SjrSrU=r $)MarkupLMOutputc(>[TU]5 [R"URUR 5Ul[R"UR URS9Ul[R"UR5Ul gr) r"r#rr%rr'rrprqr)r*r+rus r;r#MarkupLMOutput.__init__s` YYv779K9KL f&8&8f>S>STzz&"<"<= rLrrrcpURU5nURU5nURX-5nU$rrrs r;rJMarkupLMOutput.forwardrrLrrrTs@r;rrrrLrcb^\rSrSrU4SjrS\R S\R 4SjrSrU=r $)MarkupLMPooleric>[TU]5 [R"URUR5Ul[R "5Ulgr)r"r#rr%r'rTanhr-rus r;r#MarkupLMPooler.__init__s9 YYv1163E3EF '')rLrrc\USS2S4nURU5nURU5nU$)Nr)rr-)r7rfirst_token_tensor pooled_outputs r;rJMarkupLMPooler.forward s6+1a40 #56  6 rL)r-rrrTs@r;rrs($ U\\ellrLrcb^\rSrSrU4SjrS\R S\R 4SjrSrU=r $)MarkupLMPredictionHeadTransformicp>[TU]5 [R"URUR5Ul[ UR[5(a[URUl OURUl [R"URURS9Ul gr) r"r#rr%r'rrrrr transform_act_fnrprqrus r;r#(MarkupLMPredictionHeadTransform.__init__s~ YYv1163E3EF f'' - -$*6+<+<$=D !$*$5$5D !f&8&8f>S>STrLrrclURU5nURU5nURU5nU$r)rrrprs r;rJ'MarkupLMPredictionHeadTransform.forwards4 =1 --m< }5 rL)rprrrrTs@r;rrs)UU\\ellrLrc4^\rSrSrU4SjrSrSrSrU=r$)MarkupLMLMPredictionHeadi'cH>[TU]5 [U5Ul[R "UR URSS9Ul[R"[R"UR55Ul URURl g)NF)bias) r"r#r transformrr%r'ridecoder ParameterrBrrrus r;r#!MarkupLMLMPredictionHead.__init__(sm 8@yy!3!3V5F5FUS LLV->->!?@ !II rLc:URURlgr)rrr7s r; _tie_weights%MarkupLMLMPredictionHead._tie_weights5s II rLcJURU5nURU5nU$r)rrrs r;rJ MarkupLMLMPredictionHead.forward8s$}5  ]3 rL)rrr) rMrNrOrPr#rrJrRrSrTs@r;rr's &&rLrcb^\rSrSrU4SjrS\R S\R 4SjrSrU=r $)MarkupLMOnlyMLMHeadi?cB>[TU]5 [U5Ulgr)r"r#r predictionsrus r;r#MarkupLMOnlyMLMHead.__init__@s 3F;rLsequence_outputrc(URU5nU$rr)r7rprediction_scoress r;rJMarkupLMOnlyMLMHead.forwardDs ,,_=  rLrrrTs@r;rr?s(<!u||! !!rLrmodulequerykeyvalueattention_maskscalingr+ head_maskc [R"XRSS55U-n Ub"USS2SS2SS2SURS24n X-n [R R U S[RS9RUR5n [R RXURS9n UbXRSSSS5-n [R"X5n U RSS5R5n X4$)Nr r>)r@ry)ptrainingr)rBmatmul transposeshaper functionalsoftmaxfloat32toryr+rview contiguous) rrrrrrr+rkwargs attn_weights causal_mask attn_outputs r;eager_attention_forwardrJs<<}}Q':;gEL!$Q1o " o%=> #1 ==((2U]](SVVW\WbWbcL==((6??([L#nnQAq&AA ,,|3K''1-88:K  $$rLc ^\rSrSrU4SjrS S\R S\\RS\\RS\\ S\ \R 4 Sjjr S r U=r $) MarkupLMSelfAttentionifc6>[TU]5 URUR-S:wa7[ US5(d&[ SURSURS35eXlURUl[URUR- 5UlURUR-Ul [R"URUR5Ul [R"URUR5Ul [R"URUR5Ul[R"UR 5UlUR UlURS-Ulg)Nrembedding_sizezThe hidden size (z6) is not a multiple of the number of attention heads ()g)r"r#r'num_attention_headshasattr ValueErrorr8rWattention_head_size all_head_sizerr%rrrr)attention_probs_dropout_probr+attention_dropoutrrus r;r#MarkupLMSelfAttention.__init__gsD     : : :a ?PVXhHiHi#F$6$6#78 445Q8   #)#=#= #&v'9'9Frreager)r+rr)rrrrrrrrr8_attn_implementationrrrrreshaper)r7rrrrrr~ hidden_shape query_states key_states value_statesattention_interfacerroutputss r;rJMarkupLMSelfAttention.forward|s[$))#2. CCbC$*B*BC zz-055lCMMaQRS XXm,11,?II!QO zz-055lCMMaQRS (? ;; + +w 6"9$++:Z:Z"[ $7      % $}}C$2H2HLL %  % ! "));;;;FFH 1B;-JUrL) rrrr8r+rrrrrNNF)rMrNrOrPr#rBrr FloatTensorboolrrJrRrSrTs@r;rrfst607;15,1 !||!!!2!23!E--. ! $D> ! u|| !!rLrc ^\rSrSrU4SjrSrS S\RS\\RS\\RS\\ S\ \R4 S jjr S r U=r$) MarkupLMAttentionic>[TU]5 [U5Ul[ U5Ul[ 5Ulgr)r"r#rr7routputset pruned_headsrus r;r#MarkupLMAttention.__init__s0 )&1 (0 ErLc6[U5S:Xag[XRRURRUR 5up[ URRU5URl[ URRU5URl[ URRU5URl [ URRUSS9URl URR[U5- URlURRURR-URl UR RU5Ulg)Nrrr?)lenrr7rrrrrrrrrrunion)r7headsindexs r; prune_headsMarkupLMAttention.prune_headss  u:? 7 9900$))2O2OQUQbQb  -TYY__eD *499==%@ ,TYY__eD .t{{/@/@%QO )- (E(EE (R %"&))"?"?$))B_B_"_  --33E:rLrrrrrc pUR"U4UUUS.UD6nURUSU5nU4USS-nU$N)rrrrr)r7r) r7rrrrr self_outputsattention_outputrs r;rJMarkupLMAttention.forwards]yy  )/     ;;|A F#% QR(88rL)rrr7r)rMrNrOrPr#r!rBrrrrrrJrRrSrTs@r;rrsy" ;*7;15,1 ||!!2!23E--.  $D>  u|| rLrc ^\rSrSrU4SjrS S\R S\\RS\\RS\\ S\ \R 4 Sjjr S r S r U=r$) MarkupLMLayeric>[TU]5 URUlSUl[ U5Ul[ U5Ul[U5Ul g)Nr) r"r#chunk_size_feed_forward seq_len_dimr attentionr intermediaterrrus r;r#MarkupLMLayer.__init__sI '-'E'E$*6208$V, rLrrrrrc UR"U4UUUS.UD6nUSnUSSn[URURURU5n U 4U-nU$r$)r-rfeed_forward_chunkr+r,) r7rrrrrself_attention_outputsr&r layer_outputs r;rJMarkupLMLayer.forwards"& " )/ "  " 2!4(,0  # #T%A%A4CSCSUe  /G+rLcJURU5nURX!5nU$r)r.r)r7r&intermediate_outputr3s r;r1 MarkupLMLayer.feed_forward_chunks)"//0@A{{#6I rL)r-r+r.rr,r)rMrNrOrPr#rBrrrrrrJr1rRrSrTs@r;r)r)sy-7;15,1 ||!!2!23E--.  $D>  u|| 2rLr)c^\rSrSrU4Sjr\S S\RS\\RS\\RS\\ S\\ S\\ S \ \ \R\ 44S jj5rS rU=r$) MarkupLMEncoderic>[TU]5 Xl[R"[ UR 5Vs/sHn[U5PM sn5UlSUl gs snf)NF) r"r#r8rr0r1num_hidden_layersr)layergradient_checkpointing)r7r8rHr:s r;r#MarkupLMEncoder.__init__sR  ]]5IaIaCb#cCbaM&$9Cb#cd &+#$dsA%rrrroutput_hidden_states return_dictrc U(aSOSnU(aSOSn [UR5H=upU(aX4-nUbX:OSn U "SUUU US.UD6n U SnU(dM5XS4-n M? U(aX4-n[UUU S9$)N)rrrrrr)last_hidden_stater attentions) enumerater<r )r7rrrrr?r@rall_hidden_statesall_self_attentionsrH layer_modulelayer_head_mask layer_outputss r;rJMarkupLMEncoder.forwards#7BD$5b4(4OA#$58H$H!.7.CilO(+-)"3   M*!,M  &91=M & 'tn & d^&  uU\\"O3 4& & rLr9ct^\rSrSr%\\S'SrSr\S\ \ \ \ R44U4Sjj5rSrU=r$)MarkupLMPreTrainedModeli)r8markuplmc[U[R5(akURRR SUR RS9 URb%URRR5 gg[U[R5(axURRR SUR RS9 URb2URRURR5 gg[U[R5(aJURRR5 URRRS5 g[U[5(a%URRR5 gg)zInitialize the weightsr)meanstdN?)rrr%weightdatanormal_r8initializer_rangerzero_r2r\rpfill_r)r7rs r; _init_weights%MarkupLMPreTrainedModel._init_weights/s3 fbii ( ( MM   & &CT[[5R5R & S{{&   &&('  - - MM   & &CT[[5R5R & S!!- ""6#5#56<<>.  - - KK   " " $ MM   $ $S )  8 9 9 KK   " " $:rLpretrained_model_name_or_pathc,>[TU]"U/UQ70UD6$r)r"from_pretrained)clsr[ model_argsrr:s r;r]'MarkupLMPreTrainedModel.from_pretrainedAsw&'D\z\U[\\rLrB)rMrNrOrPr__annotations__base_model_prefixrY classmethodrrrosPathLiker]rRrSrTs@r;rMrM)sH "%$]HU3PRP[P[K[E\<]]]rLrMc^\rSrSrSU4SjjrSrSrSr\\ SS\ \ RS\ \ RS\ \ RS \ \ RS \ \ RS \ \ RS \ \ RS \ \ RS\ \S\ \S\ \S\\\44Sjj55rSrU=r$) MarkupLMModeliFc>[TU]U5 Xl[U5Ul[ U5UlU(a [U5OSUlUR5 g)z^ add_pooling_layer (bool, *optional*, defaults to `True`): Whether to add a pooling layer N) r"r#r8rbrr9encoderrpooler post_init)r7r8add_pooling_layerr:s r;r#MarkupLMModel.__init__IsK   ,V4&v. 0AnV,t  rLc.URR$rrrkrs r;get_input_embeddings"MarkupLMModel.get_input_embeddingsYs...rLc$XRlgrro)r7rs r;set_input_embeddings"MarkupLMModel.set_input_embeddings\s */'rLcUR5H7up#URRURR U5 M9 g)z Prunes heads of the model. heads_to_prune: dict of {layer_num: list of heads to prune in this layer} See base class PreTrainedModel N)itemsrir<r-r!)r7heads_to_pruner<rs r; _prune_headsMarkupLMModel._prune_heads_s< +002LE LL  u % / / ; ;E B3rLr[rDrErrrgrr}rr?r@rc |U bU OURRn U bU OURRn U bU OURRn UbUb [ S5eUb"UR X5 UR 5n O"UbUR 5SSn O [ S5eUb URO URn Uc[R"XS9nUc$[R"U [RU S9nURS5RS5nURURS 9nS U- S -nUbUR5S:XahURS 5RS 5RS5RS5nUR!URR"SSSS5nOCUR5S:Xa/URS5RS5RS5nUR[%UR'55RS 9nOS/URR"-nUR)UUUUUUS 9nUR+UUUU U SS9nUS nUR,bUR-U5OSn[/UUUR0UR2S9$)ao xpath_tags_seq (`torch.LongTensor` of shape `(batch_size, sequence_length, config.max_depth)`, *optional*): Tag IDs for each token in the input sequence, padded up to config.max_depth. xpath_subs_seq (`torch.LongTensor` of shape `(batch_size, sequence_length, config.max_depth)`, *optional*): Subscript IDs for each token in the input sequence, padded up to config.max_depth. Examples: ```python >>> from transformers import AutoProcessor, MarkupLMModel >>> processor = AutoProcessor.from_pretrained("microsoft/markuplm-base") >>> model = MarkupLMModel.from_pretrained("microsoft/markuplm-base") >>> html_string = " Page Title " >>> encoding = processor(html_string, return_tensors="pt") >>> outputs = model(**encoding) >>> last_hidden_states = outputs.last_hidden_state >>> list(last_hidden_states.shape) [1, 4, 768] ```NzDYou cannot specify both input_ids and inputs_embeds at the same timer>z5You have to specify either input_ids or inputs_embeds)rzrxrr)ryrRgr)r[rDrErgrr}T)rrr?r@)rC pooler_outputrrD)r8rr?use_return_dictr%warn_if_padding_and_no_attention_maskr{rzrBrrrZr|rryr@rtr;next parametersrrirjrrrD)r7r[rDrErrrgrr}rr?r@r~rzextended_attention_maskembedding_outputencoder_outputsrrs r;rJMarkupLMModel.forwardgsN2C1N-TXT_T_TqTq$8$D $++JjJj &1%cd d  "  6 6y Q#..*K  &',,.s3KTU U%.%:!! @T@T  !"ZZ CN  !"[[EJJvVN"0":":1"="G"G"J"9"<"<4::"<"N#&)@#@H"L  }}!#%//2<!>>I??))%)' + ,,  #/!5 ' *!,8< 8O O4UY )-')77&11   rL)r8rrirj)T) NNNNNNNNNNN)rMrNrOrPr#rprsrxrrrrB LongTensorrrrrrrJrRrSrTs@r;rgrgFsJ /0C1559596:59371559,0/3&*c E,,-c !!1!12c !!1!12 c !!2!23 c !!1!12 c u//0c E--.c   1 12c $D>c 'tnc d^c  u00 1c c rLrgc!^\rSrSrU4Sjr\\SS\\RS\\RS\\RS\\RS\\RS\\RS \\RS \\RS \\RS \\RS \\ S\\ S\\ S\ \ \R\ 44Sjj55rSrU=r$)MarkupLMForQuestionAnsweringic>[TU]U5 URUl[USS9Ul[ R "URUR5UlUR5 gNF)rl) r"r# num_labelsrgrNrr%r' qa_outputsrkrus r;r#%MarkupLMForQuestionAnswering.__init__sU   ++%fF ))F$6$68I8IJ rLr[rDrErrrgrr}start_positions end_positionsrr?r@rcU bU OURRn URUUUUUUUUU U SS9 nUSnURU5nUR SSS9unnUR S5R 5nUR S5R 5nSnU bU b[U R55S:aU R S5n [U R55S:aU R S5n URS5nU RSU5 U RSU5 [US9nU"UU 5nU"UU 5nUU-S - n[UUUURURS 9$) a xpath_tags_seq (`torch.LongTensor` of shape `(batch_size, sequence_length, config.max_depth)`, *optional*): Tag IDs for each token in the input sequence, padded up to config.max_depth. xpath_subs_seq (`torch.LongTensor` of shape `(batch_size, sequence_length, config.max_depth)`, *optional*): Subscript IDs for each token in the input sequence, padded up to config.max_depth. Examples: ```python >>> from transformers import AutoProcessor, MarkupLMForQuestionAnswering >>> import torch >>> processor = AutoProcessor.from_pretrained("microsoft/markuplm-base-finetuned-websrc") >>> model = MarkupLMForQuestionAnswering.from_pretrained("microsoft/markuplm-base-finetuned-websrc") >>> html_string = " My name is Niels " >>> question = "What's his name?" >>> encoding = processor(html_string, questions=question, return_tensors="pt") >>> with torch.no_grad(): ... outputs = model(**encoding) >>> answer_start_index = outputs.start_logits.argmax() >>> answer_end_index = outputs.end_logits.argmax() >>> predict_answer_tokens = encoding.input_ids[0, answer_start_index : answer_end_index + 1] >>> processor.decode(predict_answer_tokens).strip() 'Niels' ```NT rDrErrrgrr}rr?r@rrr>r?) ignore_indexr)loss start_logits end_logitsrrD)r8r|rNrsplitsqueezerrr{clamp_rrrrD)r7r[rDrErrrgrr}rrrr?r@rrlogitsrr total_loss ignored_indexloss_fct start_lossend_losss r;rJ$MarkupLMForQuestionAnswering.forwards`&1%\ 'tn\ d^\  uU\\"$@@ A\ \ rLrzC MarkupLM Model with a `token_classification` head on top. ) custom_introc^\rSrSrU4Sjr\\SS\\RS\\RS\\RS\\RS\\RS\\RS \\RS \\RS \\RS \\ S \\ S\\ S\ \ \R\ 44Sjj55rSrU=r$)MarkupLMForTokenClassificationi=cd>[TU]U5 URUl[USS9UlUR b UR O UR n[R"U5Ul [R"URUR5Ul UR5 gr)r"r#rrgrNclassifier_dropoutr*rr)r+r%r' classifierrkr7r8rr:s r;r#'MarkupLMForTokenClassification.__init__Ds   ++%fF )/)B)B)NF % %TZTnTn zz"45 ))F$6$68I8IJ rLr[rDrErrrgrr}labelsrr?r@rc lU bU OURRn URUUUUUUUUU U SS9 n U SnURU5nSnU bF[ 5nU"UR SURR 5U R S55n[UUU RU RS9$)a xpath_tags_seq (`torch.LongTensor` of shape `(batch_size, sequence_length, config.max_depth)`, *optional*): Tag IDs for each token in the input sequence, padded up to config.max_depth. xpath_subs_seq (`torch.LongTensor` of shape `(batch_size, sequence_length, config.max_depth)`, *optional*): Subscript IDs for each token in the input sequence, padded up to config.max_depth. 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]`. Examples: ```python >>> from transformers import AutoProcessor, AutoModelForTokenClassification >>> import torch >>> processor = AutoProcessor.from_pretrained("microsoft/markuplm-base") >>> processor.parse_html = False >>> model = AutoModelForTokenClassification.from_pretrained("microsoft/markuplm-base", num_labels=7) >>> nodes = ["hello", "world"] >>> xpaths = ["/html/body/div/li[1]/div/span", "/html/body/div/li[1]/div/span"] >>> node_labels = [1, 2] >>> encoding = processor(nodes=nodes, xpaths=xpaths, node_labels=node_labels, return_tensors="pt") >>> with torch.no_grad(): ... outputs = model(**encoding) >>> loss = outputs.loss >>> logits = outputs.logits ```NTrrr>rrrrD) r8r|rNrrrrrrrD)r7r[rDrErrrgrr}rrr?r@rrrrrs r;rJ&MarkupLMForTokenClassification.forwardRs\&1%L 'tnL d^L  uU\\"N2 3L L rLrz MarkupLM Model transformer with a sequence classification/regression head on top (a linear layer on top of the pooled output) e.g. for GLUE tasks. c^\rSrSrU4Sjr\\SS\\RS\\RS\\RS\\RS\\RS\\RS \\RS \\RS \\RS \\ S \\ S\\ S\ \ \R\ 44Sjj55rSrU=r$)!MarkupLMForSequenceClassificationicr>[TU]U5 URUlXl[ U5UlUR b UR O URn[R"U5Ul [R"URUR5Ul UR5 gr)r"r#rr8rgrNrr*rr)r+r%r'rrkrs r;r#*MarkupLMForSequenceClassification.__init__s   ++ %f- )/)B)B)NF % %TZTnTn zz"45 ))F$6$68I8IJ rLr[rDrErrrgrr}rrr?r@rc  U bU OURRn URUUUUUUUUU U SS9 n U SnURU5nUR U5nSnU GbURR cUR S:XaSURlOoUR S:aNU R[R:XdU R[R:XaSURlOSURlURR S:XaI[5nUR S:Xa&U"UR5U R55nOU"X5nOURR S:Xa=[5nU"URSUR 5U RS55nO,URR S:Xa[5nU"X5n[!UUU R"U R$S 9$) a  xpath_tags_seq (`torch.LongTensor` of shape `(batch_size, sequence_length, config.max_depth)`, *optional*): Tag IDs for each token in the input sequence, padded up to config.max_depth. xpath_subs_seq (`torch.LongTensor` of shape `(batch_size, sequence_length, config.max_depth)`, *optional*): Subscript IDs for each token in the input sequence, padded up to config.max_depth. labels (`torch.LongTensor` of shape `(batch_size,)`, *optional*): Labels for computing the sequence classification/regression loss. Indices should be in `[0, ..., config.num_labels - 1]`. If `config.num_labels == 1` a regression loss is computed (Mean-Square loss), If `config.num_labels > 1` a classification loss is computed (Cross-Entropy). Examples: ```python >>> from transformers import AutoProcessor, AutoModelForSequenceClassification >>> import torch >>> processor = AutoProcessor.from_pretrained("microsoft/markuplm-base") >>> model = AutoModelForSequenceClassification.from_pretrained("microsoft/markuplm-base", num_labels=7) >>> html_string = " Page Title " >>> encoding = processor(html_string, return_tensors="pt") >>> with torch.no_grad(): ... outputs = model(**encoding) >>> loss = outputs.loss >>> logits = outputs.logits ```NTrr regressionsingle_label_classificationmulti_label_classificationr>r)r8r|rNr+r problem_typerryrBrZrWr rrrrrrrD)r7r[rDrErrrgrr}rrr?r@rrrrrs r;rJ)MarkupLMForSequenceClassification.forwardsZ&1%\ 'tn\ d^\  uU\\"$<< =\ \ rLr)rrrrgrM)r)rN)?rQrdtypingrrrrBtorch.utils.checkpointrtorch.nnrrr activationsr modeling_layersr modeling_outputsr rrrrrmodeling_utilsrrrrrutilsrrrconfiguration_markuplmr get_loggerrMloggerModulerr`rbrrrrrrrrfloatrrrr)r9rMrgrrr__all__rBrLr;rsg ,, AA!9?>2   H %/ bii/ f 4 __F   299  RYY  RYY  bii$ryy0!"))!$(,% II% <<% % << % U\\* %  %% %%87BII7v* *\%.%R. bii. b]o]]8E +E E Pj #:j j Z ^ %<^  ^ B o (?o  o d rL