+ j1^RIt^RIHuHuHt^RIHt^RIH t ^RI H t H t ^RI HtHtHt^RIHtHt^RIHtHtHt^RIHt^R IHt]P8!]4t]]!R R 7!R R ]444tR .t R#)N)TorchvisionBackend) BatchFeature)group_images_by_shapereorder_images)IMAGENET_DEFAULT_MEANIMAGENET_DEFAULT_STDSizeDict) ImagesKwargsUnpack)auto_docstringis_torchdynamo_compilinglogging) TensorType)requires)backendscaa]tRt^&to^t]t]tRRRR/t RRRR/t Rt Rt Rt RtRtV3RlRltV3RlRltV3R lV3R lltR tR tR tVtV;t#)SLANeXtImageProcessorheightiwidthTc(<V^8dQhRRRS[RR/#)image torch.Tensorsizereturn)r )format __classdict__s"u/home/luca/screen-checking/.venv/lib/python3.14/site-packages/transformers/models/slanext/image_processing_slanext.py __annotate__"SLANeXtImageProcessor.__annotate__4s3@d@d@d@d  @dc VPwr4rVVPW4,WV4pVPp\VPVP 4\WV4, p\ WX,4p \ Wh,4p \P!V \PVR7p V R,\V4\V 4, ,R, p V P4P\P4p WP4, p\P!V ^8\P!V4V4p\P!V ^8\P!V 4V 4p \P!W^, 8\P !V4V4p\P!W^, 8\P"!W^, 4V 4p VR,R,P4P\P4pRV, p\P!V \PVR7pVR,\V4\V 4, ,R, pVP4P\P4pVVP4, p\P!V^8\P!V4V4p\P!V^8\P!V4V4p\P!VV^, 8\P !V4V4p\P!VV^, 8\P"!VV^, 4V4pVR,R,P4P\P4pRV, pVP%^^4P\P&4pVP\P4pV P)4pV ^,P)4pVP)4pV^,P)4pVRVR,VR,3,pVRVR,VR,3,pVRVR,VR,3,pVRVR,VR,3,p VP^V ^4p!VP^V ^4p"VP^^V 4p#VP^^V 4p$V"V$V,V#V,,,V!V$V,V#V ,,,,p%V%R,^, p%V%P%^^4P\P&4p&V&PW4W4PVP*R7#))dtypedeviceg?iNNN)r#)r%N)Nr%i )shapeviewr$maxrrroundtorcharangefloat32floatfloortoint32where zeros_like ones_like full_likeclampuint8longr#)'selfrr batch_sizechannelsrrr$scale target_height target_width target_colsrc_col src_col_floor src_col_frac weight_right weight_left target_rowsrc_row src_row_floor src_row_frac weight_bottom weight_top image_uint8 image_int32col_left col_rightrow_top row_bottompixel_top_leftpixel_top_rightpixel_bottom_leftpixel_bottom_rightweight_bottom_3d weight_top_3dweight_right_3dweight_left_3dinterpresults'&&& r_resizeSLANeXtImageProcessor._resize4s /4kk+ f :0&@DKK,s6/AAfn- U]+ \\,emmFS #e u\7J(JKcQ **5;;7 !4!4!66 {{=1#4e6F6F|6TVbc  MA$5u7G7G 7VXef {{=AI#=u|?\^jk  QY & qy(QS` %t+c188:==ekkJ \) \\-u}}VT #f m8L(LMPSS **5;;7 !4!4!66 {{=1#4e6F6F|6TVbc  MA$5u7G7G 7VXef {{=FQJ#>P\@]_kl  VaZ 'QR )SUb &,s299;>>u{{K M) kk!S),,U[[9 !nnU[[1  %%'"Q&,,. $$&#a'--/ $Q(8(7:K%KL%a)99W;M&MN':g+>@Q(QR(Jw,?7AS)ST(--aB"=!< &++Aq,?$))!Q = ^ +o.O O 1B B_WiEi i jkG$+a%((5{{:MPPW\WbWbPccr!c *<V^8dQhRS[R,RS[RS[RRRS[RS[R S[R S[R S[R S[S[S[,,R ,RS[S[S[,,R ,RS[R ,RS[R ,RS[R ,RS[S[,R ,RS[/#)rimagesr do_resizerresamplez"tvF.InterpolationMode | int | Nonedo_center_crop crop_size do_rescalerescale_factor do_normalize image_meanN image_stddo_padpad_sizedisable_groupingreturn_tensorsr)listboolr r-strrr)rrs"rrr vs0a0a^$0a0a 0a 7 0a  0a0a0a0a0aDK'$.0a4;&-0at 0aT/0a+0a j(4/!0a$ %0ar!c ,Ve&\4'g\PR4\WR7wpp/pVP 4F&wppV'dVP VVR7pVVV&K( \ VV4p\VVR7wpp/pVP 4F9wppV'dVPVV4pVPVWxWV 4pVVV&K; \ VV4pV 'dVPVWR7p\RV/VR7#)Nz&Resampling is not supported in SLANeXt)ri)rr)rhri pixel_values)data tensor_type) r logger warning_onceritemsrZr center_croprescale_and_normalizepadr)r8r]r^rr_r`rarbrcrdrerfrgrhrirjkwargsgrouped_imagesgrouped_images_indexresized_images_groupedr&stacked_imagesresized_imagesprocessed_images_groupedprocessed_imagess&&&&&&&&&&&&&&&&, r _preprocess!SLANeXtImageProcessor._preprocessvs/&  (@(B(B    H I0EV/o,,!#%3%9%9%; !E>!%N!N,: "5 )&<((>@TU0E^fv/w,,#% %3%9%9%; !E>!%!1!1.)!L!77 LV_N/= $U +&<**BDXY #xx(88xo .2B!CQ_``r!c0<V^8dQhRS[S[,/#)rrx)r r )rrs"rrr s !5r!c H<\SV`!R/VBVP4R#)N)super__init__ init_decoder)r8rx __class__s&,rrSLANeXtImageProcessor.__init__s "6" r!c ,.R OpT\^4Uu.uFpRV^, R2NK up, pT\^4Uu.uFpRV^, R2NK up, pRV9dVPR4RV9dVPR4R.V,R.,p\V4UUu/uFwr#W2bK uppVnWn.R OVnVPR,VnVPR,VnR#uupiuupiuuppi) a Initialize the decoder vocabulary for table structure recognition. Builds a character dictionary mapping HTML table structure tokens (e.g., ``, ``, ``, colspan/ rowspan attributes) to integer indices. The dictionary includes special `"sos"` (start-of-sequence) and `"eos"` (end-of-sequence) tokens. Merged `` tokens are used in place of standalone `` tokens when applicable. z colspan=""z rowspan=" soseosN) zzzzzzrz)rrr) rangeappendremove enumeratedict charactertd_tokenbos_ideos_id)r8dict_characterichars& rr"SLANeXtImageProcessor.init_decoders   %)D)QZAwa0)DD%)D)QZAwa0)DD n ,  ! !+ . ^ #  ! !& )>1UG;,5n,EF,ETW,EF '4 ii& ii& EDGsDD / Dc VVPVnVPR,p\VP4\VP4.p\VP4pVP ^R7pVP ^R7Pp.pVP^,p\V4Fp.p .p \VP^,4Fjp \WXV 3,4p V ^8d W8XdMIW9dK-VPV ,p V PV 4V PW(V 3,4Kl VPV 4\P!V 4P4P4pK .ROV^,,.RO,pRVRX/#)a Post-process the raw model outputs to decode the predicted table structure into an HTML token sequence. Converts the model's predicted probability distributions over the structure vocabulary into a sequence of HTML tokens representing the table structure. The decoded tokens are wrapped with ``, ``, and `` tags to form a complete HTML table structure. Args: outputs ([`SLANeXtForTableRecognitionOutput`]): Raw outputs from the SLANeXt model. The `last_hidden_state` field contains the predicted probability distributions over the structure vocabulary at each decoding step, with shape `(batch_size, max_text_length, num_classes)`. Returns: `dict`: A dictionary containing: - **structure** (`list[str]`): The predicted HTML table structure as a list of tokens, wrapped with ``, ``, and `
` tags. - **structure_score** (`float`): The mean confidence score across all predicted tokens. :rN)dim structurestructure_score)zzz
)z
zz)last_hidden_statepredintrrargmaxr(valuesr&rrrr*stackmeanitem)r8outputsstructure_probsignored_tokensend_idx structure_idxstructure_str_listr9 batch_indexstructure_list score_listpositionchar_idxtextrrs&& rpost_process_table_recognition4SLANeXtImageProcessor.post_process_table_recognitionsq(-- ))C.dkk*C ,<=dkk"'..1.5 )--!-4;;"((+  ,KNJ!-"5"5a"89}(-BCDarsz, 22;2EPP4FF'*   H % :VL.VLVLr # #r!