a hN@s*dZddlmZGdddeZdgZdS)z$Speech processor class for SpeechT5.)ProcessorMixincs8eZdZdZdZdZfddZddZdd ZZ S) SpeechT5Processora} Constructs a SpeechT5 processor which wraps a feature extractor and a tokenizer into a single processor. [`SpeechT5Processor`] offers all the functionalities of [`SpeechT5FeatureExtractor`] and [`SpeechT5Tokenizer`]. See the docstring of [`~SpeechT5Processor.__call__`] and [`~SpeechT5Processor.decode`] for more information. Args: feature_extractor (`SpeechT5FeatureExtractor`): An instance of [`SpeechT5FeatureExtractor`]. The feature extractor is a required input. tokenizer (`SpeechT5Tokenizer`): An instance of [`SpeechT5Tokenizer`]. The tokenizer is a required input. ZSpeechT5FeatureExtractorZSpeechT5Tokenizercst||dS)N)super__init__)selffeature_extractor tokenizer __class__l/var/www/html/assistant/venv/lib/python3.9/site-packages/transformers/models/speecht5/processing_speecht5.pyr%szSpeechT5Processor.__init__c Osn|dd}|dd}|dd}|dd}|dd}|durT|durTtd|durl|durltd|dur|dur|dur|durtd |dur|j|g|Rd|i|}n |dur|j|fi|}nd}|dur|j|||d |} | d } n*|dur*|j|fi|} | d } nd} |dur<| S| durj| |d <| d} | durj| |d<|S)a Processes audio and text input, as well as audio and text targets. You can process audio by using the argument `audio`, or process audio targets by using the argument `audio_target`. This forwards the arguments to SpeechT5FeatureExtractor's [`~SpeechT5FeatureExtractor.__call__`]. You can process text by using the argument `text`, or process text labels by using the argument `text_target`. This forwards the arguments to SpeechT5Tokenizer's [`~SpeechT5Tokenizer.__call__`]. Valid input combinations are: - `text` only - `audio` only - `text_target` only - `audio_target` only - `text` and `audio_target` - `audio` and `audio_target` - `text` and `text_target` - `audio` and `text_target` Please refer to the docstring of the above two methods for more information. audioNtext text_target audio_target sampling_ratez\Cannot process both `audio` and `text` inputs. Did you mean `audio_target` or `text_target`?z\Cannot process both `audio_target` and `text_target` inputs. Did you mean `audio` or `text`?zaYou need to specify either an `audio`, `audio_target`, `text`, or `text_target` input to process.)rr input_values input_idslabelsattention_maskdecoder_attention_mask)pop ValueErrorrrget) rargskwargsr rrrrinputstargetsrrr r r __call__(sJ              zSpeechT5Processor.__call__c Osl|dd}|dd}|dd}|dur<|dur|jj}|jj|j_|jj|g|Ri|}||j_|d}nd}|dur:|S|durh||d<| d} | durh| |d <|S) au Collates the audio and text inputs, as well as their targets, into a padded batch. Audio inputs are padded by SpeechT5FeatureExtractor's [`~SpeechT5FeatureExtractor.pad`]. Text inputs are padded by SpeechT5Tokenizer's [`~SpeechT5Tokenizer.pad`]. Valid input combinations are: - `input_ids` only - `input_values` only - `labels` only, either log-mel spectrograms or text tokens - `input_ids` and log-mel spectrogram `labels` - `input_values` and text `labels` Please refer to the docstring of the above two methods for more information. rNrrz:Cannot process both `input_values` and `input_ids` inputs.zZYou need to specify either an `input_values`, `input_ids`, or `labels` input to be padded.rr) rrrpadr isinstancelistZ feature_sizeZ num_mel_binsr) rrrrrrrrZfeature_size_hackrr r r r os@           zSpeechT5Processor.pad) __name__ __module__ __qualname____doc__Zfeature_extractor_classZtokenizer_classrrr __classcell__r r r r rs   GrN)r&Zprocessing_utilsrr__all__r r r r s