+ j{/ ^RIt^RIHtHtHt^RIHt^RIH t H t H t ^RI H t ^RIHtHtHt^RIHt]!4'd^RIt]P*!]4t!RR ] R R 7t]!RR 7]!R R] 444tR.tR#)N) AudioInput make_audio_chat_template_content make_list_of_audio_chat_template) BatchFeature)ProcessingKwargsProcessorMixinUnpack) TextInput)auto_docstringis_torch_availablelogging)requiresc >]tRt^"tRRR/RRRRRRR/R R R R R //tRtR#)AudioFlamingo3ProcessorKwargs text_kwargspaddingT audio_kwargs sampling_ratei>return_attention_mask max_length common_kwargsreturn_tensorspt padding_sideleftN)__name__ __module__ __qualname____firstlineno__ _defaults__static_attributes__r}/home/luca/screen-checking/.venv/lib/python3.14/site-packages/transformers/models/audioflamingo3/processing_audioflamingo3.pyrr"sC t   U #T |  d F  Ir#rF)total)backendsc aa]tRt^3to]tRV3Rllt]RV3RlV3Rlll4tRV3RlV3Rlllt Rt V3RlR lt V3R lR lt ] V3R lV3R ll4t] V3RlRl4tRV3RlRlltRR/RltRtV3RlRltRtVtV;t#)AudioFlamingo3Processorc z<W@nVPV4VnWPnW`n\ SV`WVR7R#)a audio_token (`Optional[str]`, *optional*, defaults to `""`): Special token used to represent audio inputs in the chat template. default_transcription_prompt (`str`, *optional*, defaults to `"Transcribe the input speech."`): Default prompt to use for transcription tasks when applying transcription requests. max_audio_len (`int`, *optional*, defaults to 600): Maximum length of audio sequences in seconds. Audio longer than this will be truncated. ) chat_templateN) audio_tokenconvert_tokens_to_idsaudio_token_iddefault_transcription_prompt max_audio_lensuper__init__)selffeature_extractor tokenizerr*r+r.r/ __class__s&&&&&&&r$r1 AudioFlamingo3Processor.__init__8s="''==kJ,H)* *]Sr#Fc <V^8dQhRS[S[S[,,RS[R,RS[R,RS[S[,RS[/#)textaudioN output_labelskwargsreturn)r listrboolr rr)format __classdict__s"r$ __annotate__$AudioFlamingo3Processor.__annotate__Ps\AA$y/)AD Ad{ A 67 A  Ar#c z<RV9d1VR,R8wd#\VPP R24hV'dRVR&\SV`!R RVRV/VBpV'dOVP R4pVR ,P 4pR Wv^8g&R WwVPP8H&WuR &\VRR 7#)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_idsr:r9mm_token_type_ids input_idslabels)data tensor_typeri) ValueErrorr5rr0__call__popcloner4 pad_token_idr) r2r9r:r;r< model_inputsrFrHr5s &&&&, r$rL AudioFlamingo3Processor.__call__Os" v %&1A*Bd*J 7 788]^_ _ 15F- .w'IeI$I&I  , 0 01D E !+.446F-1F) *<@FT^^888 9%+ "4@@r#cx<V^8dQhRS[R,RS[S[S[,,R,RS[S[,/#)r8r:Nr9r<)rr r>r r)r@rAs"r$rBrCosG d dD  d$y/)D0 d)* dr#c <\SV`!RRVRV/VBVeFVe@\V4\V48wd%\R\V4 R\V4 R24hR#R#R#)r:r9NzGot z text but z audios; they must match 1:1.r)r0validate_inputslenrK)r2r:r9r<r5s&&&,r$rT'AudioFlamingo3Processor.validate_inputsosh AeA$A&A   1c$i3u:6MtCI;jU Dabc c7N 1 r#cbV^, ^,^,pV^, ^,^,pV#)r)r2 audio_lengthsconv_output_lengthsaudio_tokens_lengthss&& r$_get_audio_token_length/AudioFlamingo3Processor._get_audio_token_lengthzs2,q0Q6: 3a 7A=A##r#c <V^8dQhRS[/#)r8r:)r)r@rAs"r$rBrCs(0(0J(0r#c \VR,VPP,4p\VPVPP,4p.p.pVFp\VP^,4p\ ^W,^, V,4p W8dD\ PRWR,, R RVP RVP R24Tp VPV 4\WV,4p \V 4F8p W,p \V ^,V,V 4p VPW|V 4K: K VPPV4pVP!V3/VBpVPR4VR&\P!\P!VR,P!R 4V4Uu.uFqP!4NK up4pVP#V4VR&.p\\%V44F'pVP'VVR 7pVPV4K) VV3#uupi) rzAudio duration (z.1fz s) exceeds zs; truncating to first zs.attention_maskinput_features_masknum_audio_tokens) audio_idx)intr3 chunk_lengthr/shapemaxloggerwarningappendminrange fetch_audiorMtorchstacksplitsumr\rUreplace_audio_token)r2r:r< window_size max_windowsper_sample_windows flat_chunksaudio_el n_samplesn_wintime_capistartend audio_inputssrYaudio_replacementsidxreplacement_texts&&, r$_process_audio&AudioFlamingo3Processor._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$C#77 PS7T   % %&6 7%/// lsIc,<V^8dQhRS[RS[RS[/#)r8rrcr=)dictrestr)r@rAs"r$rBrCs#3333SV3r#c HVR,V,pVPV,#)rb)r+)r2rrcr<rbs&&&, r$rs+AudioFlamingo3Processor.replace_audio_tokens%'(:;IF"222r#c0<V^8dQhRS[S[,/#r8r=r>r)r@rAs"r$rBrCsCC49Cr#c(<\SV`R.,#)ra)r0model_input_names)r2r5s&r$r)AudioFlamingo3Processor.model_input_namessw(,A+BBBr#c0<V^8dQhRS[S[,/#rr)r@rAs"r$rBrCs$$DI$r#c R.#)zNInput names returned always by subprocessors but not used in model's `forward`rbr)r2s&r$unused_input_names*AudioFlamingo3Processor.unused_input_namess ###r#c<V^8dQhRS[S[S[,,S[,RS[S[S[,,R,RS[S[,RS[/#)r8r:promptNr<r=)rr>rr rr)r@rAs"r$rBrCsYG G T#Y+G d3i$&G 67 G  G r#c \\V44pVUu.uFQp\V\P4'd-VP 4P 4P4MTNKS pp\V4pV^8Xd \R4hVfVP.V,pM\V\4'd V.V,pM\V\\,4'd\V4V8wd\R\V4 RV R24h.pVFVpVfVPVP4K$\V\4'dVPV4KM\R4h M \R4h\Wt4U U u.uFwrRRR R R R V /\!V 4./.NK p p p VP"!V 3R R RR RR /VB#uupiuup p i)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 [`~AudioFlamingo3Processor.apply_chat_template`] (for example `text_kwargs`, `audio_kwargs`, ...). Returns: [`BatchFeature`]: Processor outputs ready to be passed to [`AudioFlamingo3ForConditionalGeneration.generate`]. z)`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`.roleusercontenttyper9tokenizeTadd_generation_prompt return_dict)r>r isinstanceroTensordetachcpunumpyrUrKr.rtuplerk TypeErrorziprapply_chat_template) r2r:rr< audio_itemsel batch_sizepromptsitem prompt_text audio_item conversationss &&&, r$apply_transcription_request3AudioFlamingo3Processor.apply_transcription_requests2/33STY3Z.[ cnocn]_Jr5<<4P4Pryy{(..0VXXcn o% ?HI I >889JFG  $ $h+G u - -6{j( F }OJ>tDQXGYZsAc &VP!V/VB#)z)BC as previous examples used batch_decode)r)r2rr<s&*,r$ batch_decode$AudioFlamingo3Processor.batch_decode s{{D+F++r#c&<V^8dQhRS[RS[/#)r8r9r=)r)r@rAs"r$rBrCsssr#c VP4pRF8pVPV4'gKV\V4RP4pM VPR4'dVRRP4p\V4^8d7V^,VR,8Xd"V^,R9dV^RP4pV#)zY Remove the assistant prefix and surrounding quotes from a decoded transcription string. N.)z"The spoken content of the audio isz!The transcription of the audio isz!The content of the input audio isrd>"')strip startswithrUendswith)r2r9strippedprefixs&& r$r:AudioFlamingo3Processor._strip_assistant_prefix_and_quotess ::< F ""6**#CKM288:    S ! !}**,H x=A (1+""=(1+Q[B["~++-Hr#)r+r-r.r/)NzzTranscribe the input speech.iX)NF)NN)N)rrrr rvalid_processor_kwargsr1r rLrTr\rrspropertyrrrrrrr"__classdictcell__ __classcell__)r5rAs@@r$r(r(3s;T.AAA> d d$ (0(0T33CCC$$G G R  ,r#r()ro)rnp audio_utilsrrrfeature_extraction_utilsrprocessing_utilsrrr tokenization_utils_baser utilsr r r utils.import_utilsrro get_loggerrrirr(__all__rr#r$rs"ii4HH0@@*   H %$4E" :qnqqh % %r#