n ia"dZddlZddlZddlZddlZddlZddlZddlmZm Z m Z ddl Z ddl Z dZn #e$rdZYnwxYwdZdZdZd Zd Zd Zd Zd,dZdZGddeZGddZGddeeZdZGddeeZGddeZdZ GddeZ!Gdd eZ"Gd!d"eZ#Gd#d$eZ$d%Z%Gd&d'ej&Z'Gd(d)e'Z(Gd*d+e'Z)dS)-zSpecifications declare the expected variables layout of CTranslate2 models that do not load a computation graph. The model converter should make sure that each required variable of the specification is set. N)DictListOptionalTF __optional)int8 int8_float32 int8_float16 int8_bfloat16int16float16bfloat16float32)rotary_scaling_long_factorrotary_scaling_short_factorc|s|S|d|SN/)scopenames \/var/www/html/auto_sub_bot/venv/lib/python3.11/site-packages/ctranslate2/specs/model_spec.py _join_scoper)s  eeTT ""c,|dSr)splitrs r _split_scoper/s ;;s  rcrt|}|dd|d}}d||fS)Nr)rjoin)rkeysattrs r _parent_scoper$3s9   Dss)T"X4E 88E??D  rc t|jD]\}}|drt |tr;t |D]*\}}t ||t|d||fz+kt |tr!t ||t||||t|||dS)z Recursively visits a layer spec._%s_%drN) list__dict__items startswith isinstance enumerate visit_specr LayerSpec)specfnrrvalueielems rr/r/9sDM//1122 6 6 e ??3     eT " " 6$U++ T T44;ugq >Q+R+RSSSSS T y ) ) 6 ub E4(@(@ A A A A A Bt[--u 5 5 5 5 6 6rc|s|St|}|D]_} t||}#t$r?|dd\}}t||t |}Y\wxYw|S)Nr')rgetattrAttributeErrorrsplitint)r1indexr"keyr#s r index_specr>Gs    D33 34%%DD 3 3 3**S!,,KD%4&&s5zz2DDD 3 Ks*AA32A3ceZdZfdZxZS) FrozenMetacHtj|i|}d|_|S)NT)super__call___frozen)selfargskwargsinstance __class__s rrCzFrozenMeta.__call__Us+#577#T4V44r)__name__ __module__ __qualname__rC __classcell__rIs@rr@r@Ts8rr@ceZdZfdZxZS) FrozenAttrct|dr"t||std|zt||dS)NrDzAttribute %s does not exist)hasattrr9rB __setattr__)rEr=r3rIs rrSzFrozenAttr.__setattr__\s\ 4 # # FGD#,>,> F !>!DEE E C'''''r)rJrKrLrSrMrNs@rrPrP[s8(((((((((rrPc ~eZdZdZddZ ddededeeej ffd Z d Z d Z dd e eddfd ZdZdS)r0zPA layer specification declares the weights that should be set by the converters.returnNcgfd}||r%tddzdS)zVerify that the required weights are set. Raises: ValueError: If a required weight is not set in the specification. c||dSt|tjr5|jtjkr|tj}nt|tr(tjd |}nt|tr(tjd |}nSt|tr>|tkr3tj |dtj}t|tjst|tjrt#|}n0t$r)t|t&jrt+|}t-|d}t/|||dS)Nrrutf-8)dtyper )appendr-npndarrayrYfloat64astyperfloattypeboolstrOPTIONAL frombufferencodergeneric NumpyVariabletorch_is_availabletorchTensorPyTorchVariablersetattr)r1rr3 attr_nameunset_attributess r_checkz"LayerSpec.validate.._checkms} ''---%,, P;"*,,!LL44EE5)) P++0077E4(( P((--e44E3'' PH$$M%,,w*?*?rwOOOE%,, / 5"*0M0M /%e,,# / 5%,(G(G /'..$T**2.I D)U + + + + +rz/Some required model attributes are not set: %s N)_visit ValueErrorr!)rErorns @rvalidatezLayerSpec.validateest  , , , , ,4 F  C)),--.   rr%Fprefixorderedcifd}|||r1ttdSS)a!Recursively returns the weights from this layer and its children. Arguments: prefix: Prefix to prepend to all variable names. ordered: If set, an ordered list is returned instead. Returns: Dictionary mapping variables name to value. crt|tr |tkrdS|t|<dSN)r-rbrcr)r1rr3rtvars r _register_varz*LayerSpec.variables.._register_vars>%%% %8*;*;-2C FD)) * * *rc|dSNrr)xs rz%LayerSpec.variables..s !A$r)r=)rqr)sortedr+)rErtrurzrys ` @r variableszLayerSpec.variablessp 3 3 3 3 3 3 M"""  Asyy{{???@@ @ rcJ|d}t|D]|\}}|D]t\}}||krnht|\}}|sA||r,|t vr#t ||}t|||nu}dS)z4Find duplicate variables in spec and create aliases.TruN)rreversedr$ is_scalarequalSKIP_CREATING_ALIASr>rl) rErrr3 other_name other_valuerrmr1s r_alias_variableszLayerSpec._alias_variablessNN4N00 #I..  KD%+4  ' K:%%E$1#6#6 y)) K00")<<<&dE22DD)Z888E  rc5tvr,tddtfd}||dS)z-Possibly quantizes the variable of the layer.Nz7 is not a valid quantization type. Accepted types are: , cjt|tr|rdSt|d}d}t |d|z}|jdv}|rX dkr|d}tj dtj tj |z }||z}tj |}tj |tjtjjtjtjj}|tj}t'|}t'|}n dvr |d}d}t)|jdkr(|j}||jd d}tj tj |d }d ||d k<d |z }|tj|d z}tj |}|tj}|r||}t'|}t'|}ni d vr| }nO|rM dvr|d}n3 dvr|d}n dvr|d}t3||||t3|d|z|dSdS)Nr z%s_scale)rr rr ri)rr r r rr7)axisg_@)r rr)r r r )rr r)rr r )r-VariablerrrRrYtonumpyr[ramaxabsoluterintclipiinfor minmaxr^rglenshapereshape expand_dimsrrl) r1rr3r=scaleis_quantizableis_convertible old_shaper quantizations r _quantizez&LayerSpec._quantize.._quantizeseX.. %//2C2C t$$R(CE$T:+;<>>D&+DO!DLER^E1555EGENNE!LL11E 9 % i 8 8)%00E)%00EE!%GGG!HH\22E 0#>>>!HHY//EE!%BBB!HHZ00EE!%III!HHY//E D#u % % % j3.66666! r)ACCEPTED_MODEL_TYPESrrr!rq)rErrs ` rrzLayerSpec._quantizesy  # z'ModelConfig.to_dict...s@   U>>#&&    r)r*r+rEs rto_dictzModelConfig.to_dict,s1  "m1133    rc||j|<dSrx)r*)rEr=r3s r add_attributezModelConfig.add_attribute4s" crct|dd5}tj||dd|dddddS#1swxYwYdS) z'Saves the configuration as a JSON file.wrXencodingT)indent sort_keysrpN)openjsondumprwrite)rEpath config_files r save_as_jsonzModelConfig.save_as_json7s $g . . . $+ I         d # # # $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $sAA  A$'A$N)rJrKrLrrrrrrrrrr$sV..&&&    ### $ $ $ $ $rrceZdZdZdZedZedZedZdZ dde d e e d dfd Z d e d dfd Z dZdS) ModelSpecz"The top level layer specification.cF||_i|_dS)z$Initializes the model specification.N)get_default_config_config_filesrs rrzModelSpec.__init__Fs ..00  rct)z$The name of the model specification.NotImplementedErrorrs rrzModelSpec.nameK"###rcdS)zThe model specification revision. This value is incremented each time the weights layout of the model is changed (e.g. a weight is renamed). r7rrs rrevisionzModelSpec.revisionPs qrc|jS)zThe model configuration.)rrs rconfigzModelSpec.configYs |rcdS)z5Returns the default configuration used by this model.Nrrs rrzModelSpec.get_default_config^strNrfilenamerUctj|std|z|tj|}||jvrtd|z||j|<dS)z4Registers a file to be saved in the model directory.zFile %s does not existNz*A file with name %s was already registered)osrisfilerrbasenamer)rErrs r register_filezModelSpec.register_filebszw~~d## >5<== =  w''--H t{ " "IHTUU U $ Hr output_dirc|tj|d|j8|jtj|d|jD]k\}}tj||}tj|rtd|ztj ||ldS)zwSaves this model on disk. Arguments: output_dir: Output directory where the model is saved. z model.binNz config.jsonz-File %s already exists in the model directory) _serializerrr!rrrr+exists RuntimeErrorshutilcopy)rErrr destinations rsavezModelSpec.savels  Z==>>> < # L % %bgll:}&M&M N N N"k//11 + +NHd',,z8<._write_stringsm FKS[[1_==>>> FMM'22333 FKQ//00000rIrN)rr-rbrZrrrrCURRENT_BINARY_VERSIONrrrrrrY num_bytesto_bytes) rErraliasesvariablerrr3dimalias variable_namers @rrzModelSpec._serialize~s t44 + +H(1+s++ +x((((  **** $   - 1 1 1 1 1 KK C)?@@ A A A M$) $ $ $ KK C77 8 8 8 KK CY88 9 9 9( . . e d### FKS-=-=>>??? ;77CKK C 5 56666 FK->u{-K-KLLMMM FKU__->->??@@@ ENN,,---- KK CW66 7 7 7(/ - -$} e$$$ m,,,, -) - - - - - - - - - - - - - - - - - -s5HJJJrx)rJrKrLrrpropertyrrrrrbrrrrrrrrrCs,, $$X$XX%%#%#%$%%%%+s+t++++$ - - - - -rrc#K|D]H\}}t|dkr ||dfV%t|D]\}}d||dzfz|fVIdS)Nr7rr()r+rr.) vocabulariesr vocabularyr4vocabs r_flatten_vocabulariesrs(..0055j z??a   1 % % % % %%j11 5 55q1u -u44444 5 55rc VeZdZdZ d dededed eed ed ef fd ZxZS)SequenceToSequenceModelConfigz.Configuration for sequence-to-sequence models.F unk_token bos_token eos_tokendecoder_start_tokenadd_source_bosadd_source_eosc Htjd||||||d|dS)aInitializes the configuration for sequence-to-sequence models. Args: unk_token: The unknown token. bos_token: The start of sentence token. eos_token: The end of sentence token. decoder_start_token: The decoder start token. If ``None``, the token should be passed by the user in the target prefix. add_source_bos: If ``True``, ``bos_token`` will be automatically added to the source input. add_source_eos: If ``True``, ``eos_token`` will be automatically added to the source input. **kwargs: Additional configuration. )rrrrr r NrrBr) rErrrrr r rGrIs rrz&SequenceToSequenceModelConfig.__init__sP0   3))        r)rrrrFF) rJrKrLrrbrrarrMrNs@rrrs88!-2$$         &c]               rrceZdZdZfdZdZejdZejdZ de e ddfd Z de e ddfd Z d e ddfd Zdfd Zde ddffd ZxZS)SequenceToSequenceModelSpecz3Base specification for sequence to sequence models.c\tggd|_dS)z7Initializes a sequence to sequence model specification.sourcetargetN)rBr _vocabulariesrErIs rrz$SequenceToSequenceModelSpec.__init__s4   rctSrx)rrs rrz.SequenceToSequenceModelSpec.get_default_configs,...rct)z9Returns the source vocabulary size expected by the model.rrs rget_source_vocabulary_sizez6SequenceToSequenceModelSpec.get_source_vocabulary_sizerrct)z9Returns the target vocabulary size expected by the model.rrs rget_target_vocabulary_sizez6SequenceToSequenceModelSpec.get_target_vocabulary_sizerrtokensrUNcF|jd|dS)znRegisters a source vocabulary of tokens. Arguments: tokens: List of source tokens. rNrrZrErs rregister_source_vocabularyz6SequenceToSequenceModelSpec.register_source_vocabulary& 8$++F33333rcF|jd|dS)znRegisters a target vocabulary of tokens. Arguments: tokens: List of target tokens. rNrrs rregister_target_vocabularyz6SequenceToSequenceModelSpec.register_target_vocabularyrrrc2||ddS)zvRegisters a vocabulary mapping file. Arguments: path: Path to the vocabulary mapping file. zvmap.txtN)r)rErs rregister_vocabulary_mappingz7SequenceToSequenceModelSpec.register_vocabulary_mappings 4,,,,,rc t||d}|D]\}}t |t s|g}|j|}t|t|kr/td|t|t|fztt||D]P\}\}}t||kr5td| |t||fzQdS)NrzCIncorrect number of %s vocabularies: %d registered, but expected %dzK%s vocabulary %d has size %d but the model expected a vocabulary of size %d) rBrsrrr+r-r)rrrrr.zip capitalize) rEvocabulary_sizesrsizesrr4r expected_sizerIs rrsz$SequenceToSequenceModelSpec.validatesc 55775577   ,1133  KD%eT** -d3L<  CJJ.. YS..E ;<3z3SequenceToSequenceModelSpec.save..s.TTZz-a00TTTTTTrsharedrz %s_vocabulary) dictrrr)valuesallr+_save_vocabularyrBr)rErrrrr,rIs @rrz SequenceToSequenceModelSpec.saves1$2DEEFF  3 3 5 566 TTTTCSTTT T T ;$&6q&9:L(..00 I ILD& Z4)? H H H H  Z     rr)rJrKrLrrrabcabstractmethodrrrrbrr!r#rsrrMrNs@rrrsC==     /// $$$ $$$4c4t44444c4t4444------4 !s !t ! ! ! ! ! ! ! ! ! !rrc8eZdZdZ d dededeffd ZxZS) LanguageModelConfigz"Configuration for language models.rrrrrrc Btjd|||d|dS)aInitializes the configuration for language models. Args: unk_token: The unknown token. bos_token: The start of sentence token. eos_token: The end of sentence token. **kwargs: Additional configuration. )rrrNrr )rErrrrGrIs rrzLanguageModelConfig.__init__+sG          r)rrr)rJrKrLrrbrrMrNs@rr6r6(so,,!                rr6ceZdZdZfdZdZejdZde e ddfdZ d fd Z d e ddffd Z xZS) LanguageModelSpecz'Base specification for language models.cVtg|_dS)z+Initializes a language model specification.N)rBr _vocabularyrs rrzLanguageModelSpec.__init__Es' rctSrx)r6rs rrz$LanguageModelSpec.get_default_configJs"$$$rct)z2Returns the vocabulary size expected by the model.rrs rget_vocabulary_sizez%LanguageModelSpec.get_vocabulary_sizeMrrrrUNc.t||_dS)zbRegisters the vocabulary of tokens. Arguments: tokens: List of tokens. N)r)r;rs rregister_vocabularyz%LanguageModelSpec.register_vocabularyRs  <<rct|}t|j|kr&t dt|j|fzdS)NzEVocabulary has size %d but the model expected a vocabulary of size %d)rBrsr>rr;rr)rEexpected_vocabulary_sizerIs rrszLanguageModelSpec.validateZsy #'#;#;#=#= t $< < <Wt'((*BCD  = rrbr@rsrrMrNs@rr9r9Bs11 %%% $$$($s)(((((!s!t!!!!!!!!!!rr9ctj|d|z}t|dd5}t j||dddddS#1swxYwYdS)Nz%s.jsonrrXrr)r)rrr!rrr)rrrvocabulary_pathvocabulary_files rr2r2lsgll:y4/?@@O osW 5 5 55 &/!4444555555555555555555sAA"Ac~eZdZdZeejdeefdZ de fdZ eejde fdZ de ddfdZejdejfdZde fd Zejdefd Zejdefd Zejde ddfd Zejde fd ZdS)rz(Abstract base class for model variables.rUctrxrrs rrzVariable.shapevrrc2t|jdkSr|)rrrs rrzVariable.is_scalar{s4:!##rctrxrrs rrYzVariable.dtype~rrrYcF||jkr|S||Srx)rY_torErYs rrz Variable.tos$ DJ  Kxxrctrxrrs rrzVariable.numpy!###rcht|t|uo||Srx)r`_equalrEothers rrzVariable.equals*DzzT%[[(?T[[-?-??rctrxrrs rrzVariable.num_bytesrOrctrxrrs rrzVariable.to_bytesrOrctrxrrMs rrLz Variable._torOrctrxrrRs rrQzVariable._equalrOrN)rJrKrLrrr3r4rr;rrarrbrYrr[r\rrrbytesrrLrQrrrrrss22 $tCy$$$X$$4$$$$$s$$$X$   $rz$$$$@d@@@@ $3$$$$ $%$$$$ $$$$$$ $t$$$$$$rrceZdZdZdZedeefdZede fdZ de j fdZ defdZdefdZd e defd Zdefd Zd S) rgz Model variable as a Numpy array.c||_dSrxarray)rEr\s rrzNumpyVariable.__init__s  rrUc|jjSrx)r\rrs rrzNumpyVariable.shapes zrc$|jjjSrx)r\rYrrs rrYzNumpyVariable.dtypesz$$rc|jSrxr[rs rrzNumpyVariable.numpys zrc|jjSrx)r\nbytesrs rrzNumpyVariable.num_bytess z  rc4|jSrx)r\tobytesrs rrzNumpyVariable.to_bytessz!!###rrYc|dkrHtstdt|j|St j|}|j||_|S)Nrz5Converting to bfloat16 requires torch to be installed) rhrrk from_numpyr\rr[rYr^rMs rrLzNumpyVariable._tosy J  % "K#--dj99<r`r@rPr0rrrrrrr6r9r2ABCrrgrkrrrrs  ''''''''''LLL  T### !!! 6 6 6 6   ((((((((sssss jssssl    $$$$$* $$$$>[-[-[-[-[- [-[-[-|555# # # # # K# # # LU!U!U!U!U!)U!U!U!p     +   4'!'!'!'!'! '!'!'!T555*$*$*$*$*$sw*$*$*$Z+ + + + + H+ + + \1D1D1D1D1Dh1D1D1D1D1Ds 1;;