Zj5SSKrSSKJr SSKJr SSKrSSKJrJ r J r \ "5(aSSK r \ "5(aSSK r SSKJrJr SSKJr SSKJrJrJrJr SS KJrJr "S S \S S 9r"SS\S S 9r\"SS\55rS/rg)N)Path)Any)auto_docstringis_soundfile_availableis_torch_available) AudioInputmake_list_of_audio) BatchFeature) AudioKwargsProcessingKwargsProcessorMixinUnpack)PreTokenizedInput TextInputc4\rSrSr%Sr\\\4S-\S'Sr g)CsmAudioKwargs$a encoded_length_kwargs (`dict[str, Any]`, *optional*): Dictionary of keyword arguments used to compute the encoded audio sequence length. This includes parameters such as `kernel_sizes`, `strides`, `dilations`, and `use_causal_conv` that define the convolutional layers used in audio encoding. The encoded length is used to determine how many audio tokens to generate for each audio input in the text sequence. Nencoded_length_kwargs) __name__ __module__ __qualname____firstlineno____doc__dictstrr__annotations____static_attributes__rw/root/GenerationalWealth/GenerationalWealth/venv/lib/python3.13/site-packages/transformers/models/csm/processing_csm.pyrr$s S>D00r rF)totalcT\rSrSr%\\S'SSSS./SQ/SQ/S QSS .S S .S S0S.rSrg)CsmProcessorKwargs0 audio_kwargsTleftF)padding padding_sideadd_special_tokens)rrr, rr, rr,r)r,r,r,r1r,r,r,r,r,r,r-r,)r,r,r,r,r,r,r,r,r,r,r,r,r,r,r,) kernel_sizesstrides dilationsuse_causal_convi])r sampling_ratereturn_tensorspt) text_kwargsr& common_kwargsrN)rrrrrr _defaultsrrr r!r$r$0sG  ""' !QHJ#' & # +D1Ir r$c ^\rSrSrSU4Sjjr\SSj5rS\S\\ -\ \\ --S\ \ 4Sjr \SS \\-\ \-\ \-S-S\S-S \S-S \S-S\ \ 4 S jj5r\S 5rSrU=r$) CsmProcessorENc>[US5(d(SUlURUR5UlO"URUlURUl[US5(d(SUlURUR5UlO"URUlUR Ul[ TU]XUS9 g)N audio_tokenz <|AUDIO|>audio_eos_tokenz <|audio_eos|>) chat_template)hasattrrCconvert_tokens_to_idsaudio_token_idrDaudio_eos_token_idsuper__init__)selffeature_extractor tokenizerrE __class__s r!rKCsmProcessor.__init__Gs y-00*D "+"A"A$BRBR"SD (44D "+":":D y"344#2D &/&E&EdFZFZ&[D ##,#<#UnUb UbUbUcU$[XU5H|upgnUS- U-S-n Xg- n U S-n X- n XY- U -U- S-n [R"U 5S- n X-U-U - nX- nU(aU n Un OX-n X\-U -nXXUS- -- S- U-S-nM~ U$)aD Compute the length of the encoded audio sequence. Args: audio_length (int): The length of the audio sequence. kernel_sizes (list[int]): The kernel sizes for the convolutional layers. strides (list[int]): The strides for the convolutional layers. use_causal_conv (bool): Whether to use causal convolutions. r,r4)zipmathceil) audio_lengthr5r6r7r8 cur_length kernel_sizestridedilationeffective_kernel_size padding_total padding_right padding_leftn_frames ideal_length extra_paddings r!_get_encoded_length CsmProcessor._get_encoded_length]s"  7?i6G?Kb -0 -R )K%01_$@1$D !'0M)Q.M(8L":]JfTWXXHyy*Q.H#,{:]JL(5M, - - = #2]BJ$;?'CCaGFRUVVJ%.S(r audio saving_pathkwargsc z[5(d [S5e[U5n[U[[ 45(aU/nO=[U[ [45(a[SU55(d [S5e[U5[U5:wa [S5eUR"[40UD6nUSnUSn[X5Hgupx[U[R5(a,UR!5R#5R%5n[&R("XU5 Mi g)Nz/Please install `soundfile` to save audio files.c3N# UHn[U[[45v M g7fN) isinstancerr).0ps r! *CsmProcessor.save_audio..s#@qep`aAPSUY{A[A[eps#%zAInvalid input path. Please provide a string, or a list of stringsz5The number of audio and saving paths must be the samer&r9)r ImportErrorr rirrlisttupleall ValueErrorlen _merge_kwargsr$rRtorchTensorcpufloatnumpysfwrite) rLrcrdre output_kwargsr&r9 audio_valuerks r! save_audioCsmProcessor.save_audios &''OP P#5) kC; / /&-K[4-88S@qep@q=q=q`a a u:[) )TU U**   %^4 $_5 !%5NK+u||44)oo/557==? HHQ] 36r text output_labelsdepth_decoder_labels_ratioc UR"[4SURR0UD6nUSnUSnUR SS5n U S:wa"[ UR RS35e[U[5(aU/nO=[U[[45(a[SU55(d [ S 5eUV s/sHoRUR5PM n n S n Ub[U5n[!U5n [#U 5S :a/U [#U 5:wa Uc [ S 5e[ S U S U S35eUGbUR%S05n UVs/sH"oR&"UR(S40U D6PM$ nnUR+5n/nUHn/nURU;a`UR%S 5nURU-nUR-U5 UR/URSS5nURU;aM`SU;a*UR/SUR%S 5S5nSU;aM*UR-U5 M UnUR"U40UD6n0nUR1U5 UGbUR%SS5 //nnS nU GH*n U S :XaMUR-[2R4"S 55 UR-[6R8"S/55 MWUR-[2R:"UUUU -Vs/sHBn[U<5(aUR?5RA5OUPMD snSS95 UR-[6R8"UUUU -Vs/sHnUR(SPM sn5RCSS95 UU - nGM- URD"U40UD6nUR%SS5 UR1U5 [GSU55nUVs/sH>n[6RHRJRMUS UUR(S- 4SS9PM@ nn[6RN"US S9US'U(aUSURP:HRS5nUR(S n US::a.[6RT"U 5S[WU SU- -5n!UU!n"OUn"[6RX"USURP:HUSURZ:H-USS5n#SU#U"SS2S 4U"SS2S44'U#US'[]UU S9$s sn fs snfs snfs snfs snf) a output_labels (bool, *optional*, default=False): Whether to return labels for training. Indices will be in `[config.audio_token_id, -100, -101]`. - `config.audio_token_id` indicates an audio frame (considering sequence length elements as frames) - `-100` will be ignored in the loss computation - `-101` indicates the audio frame will be used only for the backbone model (using the first codebook token as labels) depth_decoder_labels_ratio (float, *optional*, default=1.0): The ratio of audio frames to keep for the depth decoder labels. Returns: [`BatchFeature`]: A [`BatchFeature`] with the following fields: - **input_ids** -- List of token ids to be fed to a model. Returned when `text` is not `None`. - **input_values** -- List of audio values to be fed to a model. Returned when `audio` is not `None`. - **attention_mask** -- List of indices specifying which tokens should be attended to by the model (when `return_attention_mask=True` or if *"attention_mask"* is in `self.model_input_names` and if `text` is not `None`). - **labels** -- List of labels for the audio frames. Returned when `output_labels=True`. tokenizer_init_kwargsr<r&r:Nr;z% only supports `return_tensors='pt'`.c3B# UHn[U[5v M g7frh)rir)rjts r!rl(CsmProcessor.__call__..s9[VZQR*Q:L:LVZszAInvalid input text. Please provide a string, or a list of stringsrz@No audio were provided, but there are audio tokens in the promptz)The number of audio tokens in each text (z7) should be the same as the number of provided audios (z).rz r,return_attention_mask)axis)dim padding_maskc3># UHoRSv M g7f)rN)shape)rjcut_idxss r!rlrsR=Q..,=Qs)valueinput_values_cutoffs input_ids?iilabels)data tensor_type)/rtr$rN init_kwargsgetrrrOrrirrorprqcountrCr rssumpoprarcopyappendreplaceupdatenpzerosrutensor concatenatervrwrycumsumrMmaxnn functionalpadstackrHnonzerorandpermintwhererIr )$rLrrcrrrer|r<r&r:rn_audio_in_textn_audior audio_arraynum_audio_tokens_listnum_audio_tokens_list_copy expanded_textsample replace_strnum_audio_tokensexpanded_audio_tokenencodingrconcatenated_audioroffsetel audio_inputsmax_lenraudio_frame_idxsn_audio_frames rand_idxsskip_frames_idxsrs$ r!__call__CsmProcessor.__call__s:**  "&.."<"<  $M2 $^4 $)94@ T ! 7 788]^_ _ dC 6DTD%=11c9[VZ9[6[6[`a a>BCd774#3#34dC  &u-E%jG  ! #33G(G} !cdd ??PQ229">  $0$4$45Lb$Q !lq%lq]h(():):2)>XBWXlq "%*?)C)C)E &M &&&0'A'E'Ea'H$+/+;+;>N+N(&&';<#^^D,<,>$6+6 H     4d ;792 4 F*a<&--bhhqk:(// bT0BC&--+0'9I*J*JB5?r5<<4P4P 0VX X*J"$ )// U6FU\L\=]%^=]rbhhrl=]%^_ffkmfng%F#+& 112DU UL   ^T 2 KK %R=QRRG!5$ 4H##''1gr@R6R2S[]'^ 4 !$,1;;7KQR+SD' (  $[ 1T5H5H HQQS -33A6N)S0!NN>:;sSSTWqSqAr=st #3I#> #3 [[k"d&9&99d;>OSWSjSj>jk[!F FJF#AqD)+;AqD+AA B#DN>BBGD$%L&_$s $U %)U A UUAU cURRnURRnUVs/sH o3S:wdM UPM nn[X-S/-5$s snf)Nrr)rNmodel_input_namesrMro)rLtokenizer_input_namesfeature_extractor_input_namesnames r!rCsmProcessor.model_input_names7sb $ @ @(,(>(>(P(P%;X(r:W$cq[q:W%(r)IMcLddee)ss AA)rDrIrCrHrh)NNNN)NFr)rrrrrK staticmethodrar rrrorr$r~rrrboolrxrpropertyrr __classcell__)rOs@r!r@r@Es  T,##J 4 44Z$sTz"22 4+, 4D$(%*36 OC++d9o=EV@WWZ^^OCD OCd{ OC %*DL OC +, OCOCbffr r@)rSpathlibrtypingrryrutilsrrrru soundfilerz audio_utilsr r feature_extraction_utilsr processing_utilsr r rrtokenization_utils_baserrrr$r@__all__rr r!rs OO94UUC 1[ 1)*yf>yfyfx  r