3j3SSKrSSKJrJr SSKJr SSKJrJ r J r SSK J r SSK JrJrJr \"5(aSSKr\R$"\5r"SS \S S 9r\"S S \ 55rS /rg)N) AudioInputmake_list_of_audio) BatchFeature)ProcessingKwargsProcessorMixinUnpack) TextInput)auto_docstringis_torch_availableloggingc2\rSrSrSS0SSSS.SSS .S .rS rg ) GlmAsrProcessorKwargs&paddingTi> max_length) sampling_ratereturn_attention_maskrptleft)return_tensors padding_side) text_kwargs audio_kwargs common_kwargsN)__name__ __module__ __qualname____firstlineno__ _defaults__static_attributes__rf/home/wildlama/miniconda3/lib/python3.13/site-packages/transformers/models/glmasr/processing_glmasr.pyrr&s2 t #%)# #"  Ir#rF)totalc ^\rSrSr\rSU4Sjjr\SS\\ \-S\ S-S\ S-S\ \S \ 4 U4S jjj5rSS\ S-S\\ \-S-S\ \4U4S jjjrSS jrS\ 4S jrS\S\S \4Sjr\S \ \4U4Sjj5r\S \ \4Sj5rSS\\ \-\ -S\\ \-S-S\ \S \ 4SjjrSS.SjrSrS\S \4SjrSrU=r$)GlmAsrProcessor7Ncr>X@lURU5UlXPlX`l[ TU]XUS9 g)aG audio_token (`Optional[str]`, *optional*, defaults to `"<|pad|>`"): Special token used to represent audio inputs in the chat template. default_transcription_prompt (`str`, *optional*, defaults to `"Please transcribe this audio into text"`): Default prompt to use for transcription tasks when applying transcription requests. max_audio_len (`int`, *optional*, defaults to 655): Maximum length of audio sequences in seconds. Audio longer than this will be truncated. 655 gives approximately 8192 tokens, corresponding to the maximum sequence length of the text model. ) chat_templateN) audio_tokenconvert_tokens_to_idsaudio_token_iddefault_transcription_prompt max_audio_lensuper__init__)selffeature_extractor tokenizerr*r+r.r/ __class__s r$r1GlmAsrProcessor.__init__;s=$''==kJ,H)* *]Sr#Ftextaudio output_labelskwargsreturnc $>SU;a+USS:wa"[URRS35eU(aSUS'[TU]"S X!S.UD6nU(a0UR S5nSXfUR R:H'XeS '[USS 9$) a output_labels (bool, *optional*, default=False): Whether to return labels for training. Returns: [`BatchFeature`]: A dictionary with tokenized text (`input_ids`, `attention_mask`) and audio features (`input_features`, `input_features_mask`). rrz% only supports `return_tensors='pt'`.Treturn_mm_token_type_idsr8r7mm_token_type_idsilabels)data tensor_typer) ValueErrorr5rr0__call__popr4 pad_token_idr)r2r7r8r9r: model_inputsr@r5s r$rDGlmAsrProcessor.__call__Ss" v %&1A*Bd*J 7 788]^_ _ 15F- .w'IeI&I !%%&9:F<@FT^^888 9%+ "4@@r#c >[TU]"SXS.UD6 UbAUb=[U5[U5:wa$[S[U5S[U5S35eggg)Nr>zGot z text but z audios; they must match 1:1.r)r0validate_inputslenrC)r2r8r7r:r5s r$rJGlmAsrProcessor.validate_inputsqsc AeA&A   1c$i3u:6MtCI;jU Dabc c7N 1 r#cdSnSHup4nUSU--US- - S- U-S-nM X- U-S-nU$)N))rrO)rOrrPrOr)r2 audio_lengths merge_factorr kernel_sizestride num_tokenss r$_get_audio_token_length'GlmAsrProcessor._get_audio_token_length|sY ,B (G&*Q[8K!OLqPU[[^__M-C$2|CaG r#c :[USURR-5n[URURR-5n/n/nUHn[URS5n[ SX-S- U-5n X:a;[ RSXS- SSURSURS35 Un URU 5 [XU-5n [U 5H,n X-n [U S-U-U 5n URX|U 5 M. M URRU5nUR"U40UD6nURS 5US '[R"[R"US R!S 5U5Vs/sHoR!5PM sn5nUR#U5US '/n[[%U55H$nUR'UUS 9nURU5 M& UU4$s snf)NrrrOzAudio duration (z.1fz s) exceeds zs; truncating to first zs.attention_maskinput_features_masknum_audio_tokens) audio_idx)intr3 chunk_lengthr/shapemaxloggerwarningappendminrange fetch_audiorEtorchstacksplitsumrVrKreplace_audio_token)r2r8r: window_size max_windowsper_sample_windows flat_chunksaudio_el n_samplesn_wintime_capistartend audio_inputssrQaudio_replacementsidxreplacement_texts r$_process_audioGlmAsrProcessor._process_audios<&1D4J4J4W4WWX $,,0F0F0S0SST (*(* HHNN1-.II3a7KGHE"&y/3J'J3&O{[_[m[mZnoFGKGYGYFZZ\]$  % %e ,9k&9:H5\1q5K/:""8##67" &&2259--kDVD .:.>.>?O.P *+  #kk,7L*M*Q*QRT*UWij kjUUWj k ,0+G+G +V '(U$C#77 PS7T   % %&6 7%/// ls"Hrxr]c0USUnURU-$)Nr\)r+)r2rxr]r\s r$rl#GlmAsrProcessor.replace_audio_tokens%'(:;IF"222r#c >[TU]S/-$)NrZ)r0model_input_names)r2r5s r$r!GlmAsrProcessor.model_input_namessw(,A+BBBr#cS/$)zNInput names returned always by subprocessors but not used in model's `forward`r\r)r2s r$unused_input_names"GlmAsrProcessor.unused_input_namess ###r#promptc ~[U[5(aU/nO[U[[45(a*U(a#[ SU55(a [U5nO[[ U55n[ 5(a]UVs/sHPn[U[R5(a,UR5R5R5OUPMR nn[U5nUS:Xa [S5eUcUR/U-nO[U[5(aU/U-nO[U[[45(a[U5U:wa[S[U5SUS35e/nUHTnUcURUR5 M#[U[5(aURU5 MK[!S5e O [!S5e[#Xt5V V s/sH-upS [U [5(aS U S .OS U S .S U S./S./PM/ n n n UR$"U 4SSSS.UD6$s snfs sn n f)a Prepare inputs for automatic speech recognition without manually writing the default transcription prompt. Args: audio (`str`, `list[str]`, `np.ndarray`, `torch.Tensor`, `list[np.ndarray]`, `list[torch.Tensor]`): Audio to transcribe. Strings are interpreted as local paths or URLs and will be loaded automatically by the chat template loader; NumPy arrays and PyTorch tensors are forwarded directly. prompt (`str` or `list[str]`, *optional*): Custom prompt(s) to include in the user turn. A list must be the same length as the batch. When `None`, each sample uses `"Transcribe the input speech."`. **kwargs: Additional keyword arguments forwarded to [`~GlmAsrProcessor.apply_chat_template`] (for example `text_kwargs`, `audio_kwargs`, ...). Returns: [`BatchFeature`]: Processor outputs ready to be passed to [`GlmAsrForConditionalGeneration.generate`]. c3B# UHn[U[5v M g7fN) isinstancestr).0els r$ >GlmAsrProcessor.apply_transcription_request..s?d^cXZ 2s@S@S^csrz)`audio` must contain at least one sample.z Received z prompt(s) for z$ audio sample(s); counts must match.z'Each prompt must be a string or `None`.z<`prompt` must be a string, a sequence of strings, or `None`.userr8)typepath)rr8r7)rr7)rolecontentT)tokenizeadd_generation_prompt return_dict)rrlisttupleallrr rhTensordetachcpunumpyrKrCr.rd TypeErrorzipapply_chat_template) r2r8rr: audio_itemsr batch_sizepromptsitem prompt_text audio_item conversationss r$apply_transcription_request+GlmAsrProcessor.apply_transcription_requests/2 eS ! !38'K e} - -%C?d^c?d889JFG  $ $h+G u . .6{j( F }OJ>tDQXGYZsAc&UR"U0UD6$)z)BC as previous examples used batch_decode)r)r2rr:s r$ batch_decodeGlmAsrProcessor.batch_decodes{{D+F++r#c`UR5nSH7nURU5(dMU[U5SR5n O URS5(aUSSR5n[U5S:a(USUS:XaUSS;aUSSR5nU$) zY Remove the assistant prefix and surrounding quotes from a decoded transcription string. )z"The spoken content of the audio isz!The transcription of the audio isz!The content of the input audio isN.r[rPr>"'rO)strip startswithrKendswith)r2r7strippedprefixs r$r2GlmAsrProcessor._strip_assistant_prefix_and_quotess ::< F ""6**#CKM288:    S ! !}**,H x=A (1+""=(1+Q[B["~++-Hr#)r+r-r.r/)Nz<|pad|>z&Please transcribe this audio into texti)NF)NN)rQ torch.Tensorr;rr) rrrr rvalid_processor_kwargsr1r r rrboolr rrDrrJrVr}dictr^rrlpropertyrrrrrrr" __classcell__)r5s@r$r'r'7s2 %MT0$(%* A$y/)AD Ad{ A ./ A  AA>$(37 dD  d$y/)D0 d)* d d(0J(0T3333C49CC$DI$$*.O T#Y+O d3i$&O ./ O  O b*/ ,ssr#r')rnp audio_utilsrrfeature_extraction_utilsrprocessing_utilsrrr tokenization_utils_baser utilsr r r rh get_loggerrrbrr'__all__rr#r$rsw,94HH0@@   H %,E"{n{{|  r#