9ihSSKrSSKrSSKrSSKrSSKrSSKJr SSKJrJ r J r SSK J r J r JrJr SSKrSSKrSSKJr SSKJr SSKJr SSKJrJr SS KJrJr SS KJrJr SS K J!r! SS K"J#r# SS K$J%r% \#"5r&\"5r'Sr(S S\R$RRS\*S\\S\\S\\+S\,S\,SS4Sjjr-S!S\S\4Sjjr.S!Sjr/S\ 4Sjr0S!S\\*\Rb4S\*S\ 4Sjjr2g)"N) OrderedDict)copytreeignore_patternsrmtree)CallableDictOptionalUnion)nn) Optimizer) _LRScheduler)File LocalStorage)ConfigJSONIteratorEncoder) ConfigFields ModelFile) copytree_py37) get_logger) is_masterc[5nUR5Hup#UR5X'M [US[55UlU$)zCopy a model state_dict to cpu. Args: state_dict (OrderedDict): Model weights on GPU. Returns: OrderedDict: Model weights on GPU. _metadata)ritemscpugetattrr) state_dictstate_dict_cpukeyvals [/var/www/html/land-doc-ocr/venv/lib/python3.13/site-packages/modelscope/utils/checkpoint.pyweights_to_cpur!sI!]N$$&!ggi' 'z; NN modelfilename optimizer lr_schedulermeta with_meta with_modelreturnc0nU(dU(d [S5eU(GanUc0nO,[U[5(d[S[ U535eSSKJn URU[R"5S9 [U[RRR5(a URn[US5(a&UR bURUR S9 XGS '[U["5(aUR%5US 'OI[U[5(a40US 'UR'5HupU R%5US U 'M Ub$[US 5(aUR%5US 'U(af[U[RRR5(a URn[)UR%55n U(dU nOXS '[*R,"5n [R."X|5 [0R2"U R55U5 SSS5 g!,(df  g=f) a:Save checkpoint to file. The checkpoint will have 3 fields: ``meta``, ``state_dict`` and ``optimizer``. By default, ``meta`` will contain version and time info. Args: model (Module): Module whose params are to be saved. filename (str): Checkpoint filename. optimizer (:obj:`Optimizer`, optional): Optimizer to be saved. lr_scheduler(:obj:`_LRScheduler`, optional): LRScheduler to be saved. meta (dict, optional): Metadata to be saved in checkpoint. with_meta (bool, optional): Save meta info. with_model(bool, optional): Save model states. z;Save meta by "with_meta=True" or model by "with_model=True"Nz%meta must be a dict or None, but got r) __version__) modelscopetimeCLASSES)r/r'r%rr&) ValueError isinstancedict TypeErrortyper-r,updater.asctimetorchr parallelDistributedDataParallelmodulehasattrr/r rrr!ioBytesIOsaverwritegetvalue) r#r$r%r&r'r(r) checkpointr,nameoptim_weightsfs r save_checkpointrF/s*J Z IK K <DD$''7T |DF F* { @ eUXX..FF G GLLE 5) $ $)B KK K .!6 i + +&/&:&:&>( # #+H:5EFGGH5tLJ * $)Y//))*[*ABIt,,!+K!8*3//*;&C,%55n6IJ:3%?demdno +< NNRS[R\]  Z '  ( (N)C D NNUV^U_`  #/z#AZz H  j) >>&" %%r"c ^^ ^^^^Sm SmSSjmU UUUU4SjnSSjmSSjmSU4Sjjn[RRU[R5n[ R "US S S 9nUb[ R"U5 U"UUUS S S 9upp[TRS 5SS5(a"U"UTRS 5R5 UU U U U S.$)a[ Load model checkpoint file and feed the parameters into the model. Args: model_to_load: The model to be load model_local_dir: The actual checkpoint dir on local disk. default_dtype: Set the default float type by 'torch.set_default_dtype' load_state_fn: An optional load_state_fn used to load state_dict into the model. Returns: c8[5nUR5HXupVXR;aJSRX/5nXdU'URU5 [ [ U5[ U/5- 5nMTXdU'MZ /n[ U5S:aUR5nXHU4$)N.r)rrjoinremovelistsetlencopy) state_dicts head_prefix!expected_keys_without_head_prefix missing_keysnew_state_dictrBr:name_with_headmissing_head_keyss r _add_head_prefix_to_state_dictBload_task_model_checkpoint.._add_head_prefix_to_state_dicts%'--/LD8!$;*=!>17~.188>#C $5^>r"cP/nU(aUHnUn U(aUSU3n O*U(a#SRURS5SS5n X;dMEXRn XRn X:wdMhURXURXR45 X M U$)NrX)rYsplitshapeappend) r_model_state_dict loaded_keysprefixadd_prefix_to_modelremove_prefix_from_modelignore_mismatched_sizesmismatched_keycheckpoint_key model_key model_shapecheckpoint_shapes r _find_mismatched_keys9load_task_model_checkpoint.._find_mismatched_keyss ""-* +#)(!N+; >@A(7!#."r"c ^^ ^ ^ /n/nTR5H`nSnSU;aURSS5nSU;aURSS5nU(dM>URU5 URU5 Mb [XV5HupTR U 5TU'M [ TSS5m TR 5mT bT Tl/m Ub U"UTUUST S9 T $S S[R4U U U U4Sjjjm T "XS 9 T $) Ngammaweightbetabiasr)rohead_prefix_keyslocal_metadata error_msgsr:c>Tc0OTRUSS05nT XS//T4nUR"U6 URR5HupEUcM T"XQU-S-5 M g)NTrX)rR_load_from_state_dict_modulesr) r:rorargsrBchildrrNmetadatars r rNMload_task_model_checkpoint.._load_state_dict_into_model..loadsv'/'7X\\3BK>%"FD"b"$,,d3#)??#8#8#:KD(UTMC$78$;r")ro)) keysreplacerlzippoprr^rr Module) r#r start_prefixr load_state_fnold_keysnew_keysrnew_keyold_keyrrNrs ` @@@r _load_state_dict_into_model?load_task_model_checkpoint.._load_state_dict_into_models??$CG#~++gx8}++ff5w$(%!$H 7 G",.."9Jw !8:{D9__&  #+J   $ #!1#%  '* 9RYY 9 9  ,r"c>^*^+^,UR5n[UR55n[UR55nURm+[ T+5S:a+[ U+4SjU55n[ U+4SjU55n OSnSn U(+=(a U n U=(a U (+n U (avUV s/sHoR T+5(aMU PM sn m*UV s/sH>n U R T+5(a#SRU RS5SS5OU PM@ nn O,U (a%UV s/sHn SRT+U /5PM nn /m*[[U5[U5- 5n [[U5[U5- 5nURn/n/n/n[5n[5n[ U5S:a[ U5S:a[U[5(aU/nUHm,[ U,4SjU55UT,'M UHdm,UV s/sH>n U R T,5(dMSRU RS5SS5PM@ nn URU5 UUT,'Mf [[U5[U5- 5n[[U5[U5- 5nT0RSS5nT0RS S5nUb5UH/nU Vs/sHn[ R""UU5bMUPM! n nM1 Ub5UH/nUVs/sHn[ R""UU5bMUPM! nnM1 U(a$T1"UU T+U U S 9nUHnUR%U5 M [5n[ U5S:a,UH&nUU(aMT-"UUUUU 5unnn UUU'M( S nUn[5n [ UR5S:a1['UUR5(dU(aURS-n[ UR5S:at['UUR5(aYU(dR[)XR5nUHn[)UU5U U'M [ U*4S jU55(a [+S 5eT."UUUT+U U U5n!T/"UUUU5n"[ U 5S:a+U H%n#T/"U U#UU#S-U5n$U"RU$5 M' [ U"5S:a6SRU"5n%[-SUR.R0SU%35e[ U5S:a_[2R5SUR.R0SUSUR.R0SUR.R0S3 5 O[ U5S:aH[2R5SUR.R0SUSUR.R0S35 O-[2R7SUR.R0S35 [ U 5S:a1[2R5SUR.R0SU S35 OS[ U!5S:XaD[2R7SUR.R0SUR.R0S35 [ U!5S:adS RU!V&V'V(s/sHun&n'n(S!U&S"U'S#U(S$3PM sn(n'n&5n)[2R5SUR.R0S%U)S35 XU!U"4$s sn fs sn fs sn fs sn fs snfs snfs sn(n'n&f)&Nrc3D># UHoRT5v M g7fN startswith.0sros r Gload_task_model_checkpoint.._load_checkpoint..'s /D.B V$$.B c3D># UHoRT5v M g7frrrs r rr)s.=.; V$$mrFrXric3D># UHoRT5v M g7frr)rr prefix_heads r rrRs >M7K!LL--7Kr_keys_to_ignore_on_load_missing"_keys_to_ignore_on_load_unexpected)ro add_prefix remove_prefixrc3.># UH nUT;v M g7fr)rr%expected_keys_not_base_model_prefixeds r rrs42s??2szjThe state dictionary of the model you are trying to load is corrupted. Are you sure it was properly saved?z z#Error(s) in loading state_dict for z: zESome weights of the model checkpoint were not used when initializing z: z, - This IS expected if you are initializing z from the checkpoint of a model trained on another task or with another architecture (e.g. initializing a BertForTokenClassification model from a BertForPreTraining model). - This IS NOT expected if you are initializing z from the checkpoint of a model that you expect to be exactly identical (initializing a BertForTokenClassification model from a BertForTokenClassification model).z0 - This IS Not expected if you are initializing z from the checkpoint of a model with a same task while the structure is different (e.g. initializing a BertForTokenClassification model from a BertForTokenClassification model).z9All model checkpoint weights were used when initializing z. zSome weights of zK were not initialized from the model checkpoint and are newly initialized: zo You should probably TRAIN this model on a down-stream task to be able to use it for predictions and inference.zAll the weights of z were initialized from the model checkpoint If your task is similar to the task the model of the checkpoint was trained on, you can already use z* for predictions without further training. z- z: found shape z in the checkpoint and z in the model instantiatedzl were not initialized from the model checkpoint and are newly initialized because the shapes did not match: )rr[rbase_model_prefixr]anyrrYrjr\r`r2r1strextendrresearch _init_weightsr;rr0 RuntimeError __class____name__rPrQinfo)2r#rrrr _fast_initrm expected_keyskeys_from_pretrainedpretrained_has_prefix_modulemodel_expects_prefix_modulerqrprrbunexpected_keys prefix_heads&expected_head_keys_without_head_prefixreunexpected_head_keyspretrained_has_prefix_headrrarrpatkuninitialized_modulesr:missing_head_keys_by_headr`r model_to_load heads_to_loadmismatched_keysrheadlocal_error_msgs error_msgrshape1shape2mismatched_warningrrorrfrxrkwargsretrieve_modules_from_namess2 @@@r _load_checkpoint4load_task_model_checkpoint.._load_checkpoints!++--2245 #JOO$56(( v;?+./D.B/D,D (*-.=.;.=+= ',1 (*/ ''C#C#cHc :^C^?^ #(5(a V0D=5 1 '&A/0ll6.B.Bab)*I& M!M7K>M;M*;7 , , 8E58E1||K0/CHHQWWS\!"-. 257==57$E!! ,$(O$<=>$? #O$<=>?O+1** -t+5'-3ZZ 0$.8* + 66+ +!ryya/@A|  7 . 99.#.!"))C2CA #: $?.6 %8 ! 0##F+0%)F!  1 $/ 1+>>B`"K1A+1N$C&?J 1<>O-k: 0   u&& '! +G''5)5)-I 22S8L u&& '! +''1)1)1M#E+B+BCM+ -4UK-H k* ,42444 &'' 0     $ # 1 1=}N  }  !%#>!$'TCZ$P !!"23& z?Q  J/I5eoo6N6N5OuU^T_`   ! # NN!!&!9!9 :"_&>%?@88=8P8P7QR   ! #!%+:,+:'CSEx/FvhNhi+:,"  NN"5??#;#;"<=./0<< = ozIIc5 K25.  #P,sC<\?\?&A]9]  ]>']1]])] ]'] Fc [UVs/sH&nSRURS5SS5PM( sn5nUR[UVs/sH@nUSR 5(dMSRURS5SS5PMB sn55n/nUR 5HupU(a<UR U5(a#SRURS5SS5OUnO*U(a#[U5S:aSRX(/5OUnX;dMyURU 5 M U$s snfs snf)NrXrrir) r\rYrjunionisdigit named_modulesrr]rl) r#namesrorrr module_keysretrieved_modulesrBr:s r r?load_task_model_checkpoint..retrieve_modules_from_namess. EJES388CIIcN3B$78EJK "'' 9>9>#r7??$.3,-  !//1LD,0OOF,C,CxxJJsOAB')IM36t9q=sxx/f"!((02! -K s-EE 2'E cU(a4[R"URR55UlOURUl[ USS5bx[R R URRSURRSURRS- 4SS5URl[US5(a$[US5(aURUl ggg)Nr~rconstant out_featuresnum_embeddings) r Parameterr|cloner functionalpadr~datarkr;rr)output_embeddingsinput_embeddings torchscripts r _tie_or_clone_weights9load_task_model_checkpoint.._tie_or_clone_weightss ')|| ''--/(1  $(8'>'>  $ $fd 3 ?*,--*;*;!&&++%,,2215',,22156  +  " " ' $n 5 5' "2;4;4-=-L-L  *;4 5r"c>U(aAURR5nUb#URR5nT"X#5 gggr)rget_output_embeddingsencoderget_input_embeddings)r#tie_word_embeddingsrrrs r tie_weights/load_task_model_checkpoint..tie_weightssE  % @ @ B  ,#(==#E#E#G %&7J- r"rTrH)rrrrrr)r#rbrrrr)NFF)F) rKrLrYrTORCH_MODEL_BIN_FILEr7rNset_default_dtyperrRr)rmodel_local_dir default_dtyperrrr ckpt_filerrbrrrrfrxrrrs ` @@@@@r load_task_model_checkpointrs"?$F 4lOJOJf,0/427 !>+0M2K _i.L.LMIIEMJ   .AQ# $ B>L?vzz&!#8%@@M6::f#5#I#IJ$**  r"configc[U[5(aUR5n[RU;a&SU[R 0U[R'[ R"US[S9n[RRU[R5n[RUR!5U5 g)Nr4)indentcls)r1rto_dictrpipelinetaskjsondumpsrrKrLrYr CONFIGURATIONstorager?encode) target_folderrcfg_str config_files r save_configurationr>s&&!!!F*)/ 8I8I1J(K|$$%jj/BCG'',,}i.E.EFK MM'.."K0r"rsave_checkpoint_name save_functionc Ub[U[5(d [S5eUb$[RR U5(a[ SUS35eUc [S5e[RRX5nU/n[U5nUR[R5 URS5 URS5 [US5(a`URbS[5(aD[R R"S :a[$nO[&nU"URU[)U6S S 9 U"X40UD6 g![a*n [S [+U 5R,S U S35eSn A ff=f)asave the pretrained model, its configuration and other related files to a directory, so that it can be re-loaded Args: model (Model): Model whose params are to be saved. target_folder (Union[str, os.PathLike]): Directory to which to save. Will be created if it doesn't exist. save_checkpoint_name (str): The checkpoint name to be saved in the target_folder save_function (Callable): The function to use to save the state dictionary. Nz'A valid save function must be passed inzProvided path (z#) should be a directory, not a filez6At least pass in one checkpoint name for saving methodz *.safetensorsz.* model_dirT)ignore dirs_exist_okz)During saving checkpoints, the error of "z with msg z thrown)r1r ExceptionrKrLisfiler0rYr\addrrr;r rsys version_infominorrrrr4r) r#rr r routput_ckpt_pathorigin_file_to_be_ignoredignore_file_set copytree_funces r save_pretrainedrHsy(J}h$G$GABB} = =m_,O P  # DF Fww||MH"6 634O //0(u % ;    ! !Q &$M)M OO "O4  $e88 $7Q8H8H7IJs' #$ $$s E F%FF)NNNTT)NN)3r<rKrrr. collectionsrshutilrrrtypingrrr r rr7r torch.optimr torch.optim.lr_schedulerr modelscope.fileiorrmodelscope.utils.configrrmodelscope.utils.constantrrmodelscope.utils.file_utilsrmodelscope.utils.loggerrmodelscope.utils.torch_utilsrrPrr!rrr2boolrFrUrrPathLikerrr"r r)sH #4422 !10?=5.2  .&6:;?+/&*'+ D+588??D+!D+' 2D+#+<"8D+#4. D+ $ D+ !% D+ 15 D+R,015&&(&&#/&&V.2-1\~ 1d115.2;$#(bkk)9#:;$*-;$$,;$r"