`i dZddlZddlmZddlmZmZddlmZm Z m Z ddl m Z m Z mZdd lmZdd lmZGd d ed ZGdde ZdgZdS)z% Speech processor class for Wav2Vec2 N)contextmanager)OptionalUnion)ProcessingKwargsProcessorMixinUnpack) AudioInputPreTokenizedInput TextInput)Wav2Vec2FeatureExtractor)Wav2Vec2CTCTokenizerceZdZiZdS)Wav2Vec2ProcessorKwargsN)__name__ __module__ __qualname__ _defaults/home/jaya/work/projects/VOICE-AGENT/VIET/agent-env/lib/python3.11/site-packages/transformers/models/wav2vec2/processing_wav2vec2.pyrrsIIIrrF)totalc eZdZdZdZdZfdZefdZ dde e de e e e e eefd eefd Zd Zed Zed ZxZS)Wav2Vec2Processora Constructs a Wav2Vec2 processor which wraps a Wav2Vec2 feature extractor and a Wav2Vec2 CTC tokenizer into a single processor. [`Wav2Vec2Processor`] offers all the functionalities of [`Wav2Vec2FeatureExtractor`] and [`PreTrainedTokenizer`]. See the docstring of [`~Wav2Vec2Processor.__call__`] and [`~Wav2Vec2Processor.decode`] for more information. Args: feature_extractor (`Wav2Vec2FeatureExtractor`): An instance of [`Wav2Vec2FeatureExtractor`]. The feature extractor is a required input. tokenizer ([`PreTrainedTokenizer`]): An instance of [`PreTrainedTokenizer`]. The tokenizer is a required input. r AutoTokenizercrt|||j|_d|_dS)NF)super__init__feature_extractorcurrent_processor_in_target_context_manager)selfr tokenizer __class__s rrzWav2Vec2Processor.__init__3s7 *I666!%!7*/'''rc  tj|fi|S#ttf$rWt jd|jdttj|fi|}tj|fi|}|||cYSwxYw)NzLoading a tokenizer inside a from a config that does not include a `tokenizer_class` attribute is deprecated and will be removed in v5. Please add `'tokenizer_class': 'Wav2Vec2CTCTokenizer'` attribute to either your `config.json` or `tokenizer_config.json` file to suppress this warning: )r r$) rfrom_pretrainedOSError ValueErrorwarningswarnr FutureWarningrr)clspretrained_model_name_or_pathkwargsr r$r%s rr'z!Wav2Vec2Processor.from_pretrained8s Q*577*+HSSFSS S$ Q Q Q M2cl222     !9 HIf q qjp q q ,<=Zee^deeI3):iPPP P P P QsA%BBNaudiotextr/c d|vr)tjd|d}||td|jt fd|jji|}|jr$|j |fi|d|d|dS||j |fi|d}||j|fi|d}||S||S|d |d <|S) a* This method forwards all arguments to [`Wav2Vec2FeatureExtractor.__call__`] and/or [`PreTrainedTokenizer.__call__`] depending on the input modality and returns their outputs. If both modalities are passed, [`Wav2Vec2FeatureExtractor.__call__`] and [`PreTrainedTokenizer.__call__`] are called. Args: audio (`np.ndarray`, `torch.Tensor`, `List[np.ndarray]`, `List[torch.Tensor]`, *optional*): An audio input is passed to [`Wav2Vec2FeatureExtractor.__call__`]. text (`str`, `List[str]`, *optional*): A text input is passed to [`PreTrainedTokenizer.__call__`]. Returns: This method returns the results of each `call` method. If both are used, the output is a dictionary containing the results of both. raw_speechzLUsing `raw_speech` as a keyword argument is deprecated. Use `audio` instead.NzAYou need to specify either an `audio` or `text` input to process.tokenizer_init_kwargs audio_kwargs text_kwargs common_kwargs input_idslabels) r*r+popr) _merge_kwargsrr$ init_kwargsr"r!r ) r#r0r1imagesvideosr/ output_kwargsinputs encodingss r__call__zWav2Vec2Processor.__call__KsG, 6 ! ! Mh i i iJJ|,,E =T\`aa a** #  "&."<    * )4)/ . 0    +T+ESS]>5RSSF  &tLL}]/KLLI <M ] (5F8 MrcZ|jr|jj|i|S|dd}|dd}t |dkr|d}|dd}||jj|g|Ri|}||jj|fi|}||S||S|d|d<|S)ag This method operates on batches of extracted features and/or tokenized text. It forwards all arguments to [`Wav2Vec2FeatureExtractor.pad`] and/or [`PreTrainedTokenizer.pad`] depending on the input modality and returns their outputs. If both modalities are passed, [`Wav2Vec2FeatureExtractor.pad`] and [`PreTrainedTokenizer.pad`] are called. Args: input_features: When the first argument is a dictionary containing a batch of tensors, or the `input_features` argument is present, it is passed to [`Wav2Vec2FeatureExtractor.pad`]. labels: When the `label` argument is present, it is passed to [`PreTrainedTokenizer.pad`]. Returns: This method returns the results of each `pad` method. If both are used, the output is a dictionary containing the results of both. input_featuresNr9rr r8)r"r!padr:lenr r$)r#argsr/rDr9s rrEzWav2Vec2Processor.pads  * ?-4)-t>v>> >$4d;;Hd++ t99q==!!WN8D  %7T37XXXXQWXXN  'T^'99&99F >! !  #M'-k':N8 $! !rc&|jj}|dgzS)Nr9)r model_input_names)r#feature_extractor_input_namess rrIz#Wav2Vec2Processor.model_input_namess)-(>(P%,z99rc#Ktjdd|_|j|_dV|j|_d|_dS)z Temporarily sets the tokenizer for processing the input. Useful for encoding the labels when fine-tuning Wav2Vec2. z`as_target_processor` is deprecated and will be removed in v5 of Transformers. You can process your labels by using the argument `text` of the regular `__call__` method (either in the same call as your audio inputs, or in a separate call.TNF)r*r+r"r$r!r )r#s ras_target_processorz%Wav2Vec2Processor.as_target_processorsW   8   +/'!% !%!7*/'''r)NNNN)rrr__doc__feature_extractor_classtokenizer_classr classmethodr'rr rstrlistr r r rrBrEpropertyrIrrL __classcell__)r%s@rrr!s&  9%O00000 QQQQ[Q('+NR 66 #6uS$s)Y8IIJK6 01 6666p#"#"#"J::X: 00^00000rr)rMr* contextlibrtypingrrprocessing_utilsrrr tokenization_utils_baser r r feature_extraction_wav2vec2rtokenization_wav2vec2rrr__all__rrrr\s%%%%%%""""""""HHHHHHHHHHOOOOOOOOOOAAAAAA777777.e\0\0\0\0\0\0\0\0~  r