§
    9²¸jÜX  ã                   ó¤   — d Z ddlmZ ddlZddlmZmZmZm	Z	 ddl
mZ ddlmZ ddlmZmZmZ  ej        e¦  «        Z G d	„ d
e¦  «        Zd
gZdS )z)Feature extractor class for UnivNetModel.é    )ÚAnyNé   )Úmel_filter_bankÚoptimal_fft_lengthÚspectrogramÚwindow_function)ÚSequenceFeatureExtractor)ÚBatchFeature)ÚPaddingStrategyÚ
TensorTypeÚloggingc            )       ój  ‡ — e Zd ZdZg d¢Z	 	 	 	 	 	 	 	 	 	 	 	 	 	 	 	 	 	 	 	 	 d>dedededededededededz  dedededz  d ed!ed"ed#ed$ed%ed&ed'ef(ˆ fd(„Z	d)„ Z
d*„ Zd+ej        d,ej        fd-„Z	 d?d.ed/ej        j        dz  d,ej        fd0„Zd?d,eej                 fd1„Z	 	 	 	 	 	 	 	 	 	 	 	 d@d2ej        ee         z  eej                 z  eee                  z  dedz  d3eez  ez  d4edz  d5ed6edz  d7ed/ej        j        dz  d8ed9edz  dedz  d:edz  d;eez  dz  d,efd<„Zd,eeef         fˆ fd=„Zˆ xZS )AÚUnivNetFeatureExtractoraì  
    Constructs a UnivNet feature extractor.

    This class extracts log-mel-filter bank features from raw speech using the short time Fourier Transform (STFT). The
    STFT implementation follows that of TacoTron 2 and Hifi-GAN.

    This feature extractor inherits from [`~feature_extraction_sequence_utils.SequenceFeatureExtractor`] which contains
    most of the main methods. Users should refer to this superclass for more information regarding those methods.

    Args:
        feature_size (`int`, *optional*, defaults to 1):
            The feature dimension of the extracted features.
        sampling_rate (`int`, *optional*, defaults to 24000):
            The sampling rate at which the audio files should be digitalized expressed in hertz (Hz).
        padding_value (`float`, *optional*, defaults to 0.0):
            The value to pad with when applying the padding strategy defined by the `padding` argument to
            [`UnivNetFeatureExtractor.__call__`]. Should correspond to audio silence. The `pad_end` argument to
            `__call__` will also use this padding value.
        do_normalize (`bool`, *optional*, defaults to `False`):
            Whether to perform Tacotron 2 normalization on the input. Normalizing can help to significantly improve the
            performance for some models.
        num_mel_bins (`int`, *optional*, defaults to 100):
            The number of mel-frequency bins in the extracted spectrogram features. This should match
            `UnivNetModel.config.num_mel_bins`.
        hop_length (`int`, *optional*, defaults to 256):
            The direct number of samples between sliding windows. Otherwise referred to as "shift" in many papers. Note
            that this is different from other audio feature extractors such as [`SpeechT5FeatureExtractor`] which take
            the `hop_length` in ms.
        win_length (`int`, *optional*, defaults to 1024):
            The direct number of samples for each sliding window. Note that this is different from other audio feature
            extractors such as [`SpeechT5FeatureExtractor`] which take the `win_length` in ms.
        win_function (`str`, *optional*, defaults to `"hann_window"`):
            Name for the window function used for windowing, must be accessible via `torch.{win_function}`
        filter_length (`int`, *optional*, defaults to 1024):
            The number of FFT components to use. If `None`, this is determined using
            `transformers.audio_utils.optimal_fft_length`.
        max_length_s (`int`, *optional*, defaults to 10):
            The maximum input length of the model in seconds. This is used to pad the audio.
        fmin (`float`, *optional*, defaults to 0.0):
            Minimum mel frequency in Hz.
        fmax (`float`, *optional*):
            Maximum mel frequency in Hz. If not set, defaults to `sampling_rate / 2`.
        mel_floor (`float`, *optional*, defaults to 1e-09):
            Minimum value of mel frequency banks. Note that the way [`UnivNetFeatureExtractor`] uses `mel_floor` is
            different than in [`transformers.audio_utils.spectrogram`].
        center (`bool`, *optional*, defaults to `False`):
            Whether to pad the waveform so that frame `t` is centered around time `t * hop_length`. If `False`, frame
            `t` will start at time `t * hop_length`.
        compression_factor (`float`, *optional*, defaults to 1.0):
            The multiplicative compression factor for dynamic range compression during spectral normalization.
        compression_clip_val (`float`, *optional*, defaults to 1e-05):
            The clip value applied to the waveform before applying dynamic range compression during spectral
            normalization.
        normalize_min (`float`, *optional*, defaults to -11.512925148010254):
            The min value used for Tacotron 2-style linear normalization. The default is the original value from the
            Tacotron 2 implementation.
        normalize_max (`float`, *optional*, defaults to 2.3143386840820312):
            The max value used for Tacotron 2-style linear normalization. The default is the original value from the
            Tacotron 2 implementation.
        model_in_channels (`int`, *optional*, defaults to 64):
            The number of input channels to the [`UnivNetModel`] model. This should match
            `UnivNetModel.config.model_in_channels`.
        pad_end_length (`int`, *optional*, defaults to 10):
            If padding the end of each waveform, the number of spectrogram frames worth of samples to append. The
            number of appended samples will be `pad_end_length * hop_length`.
        return_attention_mask (`bool`, *optional*, defaults to `True`):
            Whether or not [`~UnivNetFeatureExtractor.__call__`] should return `attention_mask`.
    )Úinput_featuresÚnoise_sequenceÚpadding_maské   éÀ]  ç        Féd   é   é   Úhann_windowé
   Nç•Ö&è.>ç      ð?çñhãˆµøä>ç    ž'Àç    Äƒ@é@   TÚfeature_sizeÚsampling_rateÚpadding_valueÚdo_normalizeÚnum_mel_binsÚ
hop_lengthÚ
win_lengthÚwin_functionÚfilter_lengthÚmax_length_sÚfminÚfmaxÚ	mel_floorÚcenterÚcompression_factorÚcompression_clip_valÚnormalize_minÚnormalize_maxÚmodel_in_channelsÚpad_end_lengthc           	      ó–  •—  t          ¦   «         j        d||||dœ|¤Ž || _        || _        || _        || _        || _        |	| _        || _        |€t          |¦  «        dz  }|| _
        || _        |
| _        |
|z  | _        | j        €t          | j        ¦  «        | _        n| j        | _        | j        dz  dz   | _        t#          | j        | j        d¬¦  «        | _        t'          | j        | j        | j        | j
        | j        dd¬¦  «        | _        || _        || _        || _        || _        || _        || _        || _        d S )	N)r!   r"   r#   Úreturn_attention_maské   r   T)Úwindow_lengthÚnameÚperiodicÚslaney)Únum_frequency_binsÚnum_mel_filtersÚmin_frequencyÚmax_frequencyr"   ÚnormÚ	mel_scale© )ÚsuperÚ__init__r$   r%   r&   r'   r(   r)   r+   Úfloatr,   r-   r*   Únum_max_samplesr   Ún_fftÚn_freqsr   Úwindowr   r"   Úmel_filtersr.   r/   r0   r1   r2   r3   r4   )Úselfr!   r"   r#   r$   r%   r&   r'   r(   r)   r*   r+   r,   r-   r.   r/   r0   r1   r2   r3   r4   r6   ÚkwargsÚ	__class__s                          €ú†/root/private/project/pallet_detection_v2/.venv/lib/python3.11/site-packages/transformers/models/univnet/feature_extraction_univnet.pyrD   z UnivNetFeatureExtractor.__init__e   s|  ø€ ð2 	�‰ŒÔð 	
Ø%Ø'Ø'Ø"7ð		
ð 	
ð
 ð	
ð 	
ð 	
ð )ˆÔà(ˆÔØ$ˆŒØ$ˆŒØ(ˆÔØ*ˆÔØˆŒ	Øˆ<å˜Ñ'Ô'¨!Ñ+ˆDØˆŒ	Ø"ˆŒà(ˆÔØ+¨mÑ;ˆÔàÔÐ%Ý+¨D¬OÑ<Ô<ˆDŒJˆJàÔ+ˆDŒJØœ
 a™¨1Ñ,ˆŒå%°D´OÈ$ÔJ[ÐfjÐkÑkÔkˆŒå*Ø#œ|Ø Ô-Øœ)Øœ)ØÔ,ØØð
ñ 
ô 
ˆÔð ˆŒØ"4ˆÔØ$8ˆÔ!Ø*ˆÔØ*ˆÔØ!2ˆÔØ,ˆÔÐÐó    c                 óB   — d|| j         z
  | j        | j         z
  z  z  dz
  S )Nr7   r   ©r1   r2   ©rK   r   s     rN   Ú	normalizez!UnivNetFeatureExtractor.normalize±   s+   € Ø�[ 4Ô#5Ñ5¸$Ô:LÈtÔOaÑ:aÑbÑcÐfgÑgÐgrO   c                 óB   — | j         | j        | j         z
  |dz   dz  z  z   S )Nr   r7   rQ   rR   s     rN   Údenormalizez#UnivNetFeatureExtractor.denormalize´   s-   € ØÔ! TÔ%7¸$Ô:LÑ%LÐR]Ð`aÑRaÐefÑQfÑ$gÑgÐgrO   ÚwaveformÚreturnc                 óT  — t          j        |t          | j        | j        z
  dz  ¦  «        t          | j        | j        z
  dz  ¦  «        fd¬¦  «        }t          || j        | j        | j        | j        d| j        dd¬¦	  «	        }t          j        t          j	        |¦  «        dz  t          j
        |¦  «        dz  z   | j        z   ¦  «        }t          j        | j        j        |¦  «        }t          j        t          j        || j        d¬¦  «        | j        z  ¦  «        }|j        S )aú  
        Calculates log MEL spectrograms from a batch of waveforms. Note that the input waveform(s) will be padded by
        `int(self.n_fft - self.hop_length) / 2` on both sides using the `reflect` padding mode.

        Args:
            waveform (`np.ndarray` of shape `(length,)`):
                The input waveform. This must be a single real-valued, mono waveform.

        Returns:
            `numpy.ndarray`: Array containing a log-mel spectrogram of shape `(num_frames, num_mel_bins)`.
        r7   Úreflect)ÚmodeN)rI   Úframe_lengthr&   Ú
fft_lengthÚpowerr.   rJ   r-   )Úa_minÚa_max)ÚnpÚpadÚintrG   r&   r   rI   r.   ÚsqrtÚrealÚimagr-   ÚmatmulrJ   ÚTÚlogÚclipr0   r/   )rK   rV   Úcomplex_spectrogramÚamplitude_spectrogramÚmel_spectrogramÚlog_mel_spectrograms         rN   rl   z'UnivNetFeatureExtractor.mel_spectrogram·   s0  € õ ”6ØÝ�$”*˜tœÑ.°!Ñ3Ñ4Ô4µc¸4¼:ÈÌÑ;WÐ[\Ñ:\Ñ6]Ô6]Ð^Øð
ñ 
ô 
ˆõ *ØØ”;ØœØ”Ø”zØØ”;ØØð

ñ 

ô 

Ðõ !#¤ÝŒGÐ'Ñ(Ô(¨AÑ-µ´Ð8KÑ0LÔ0LÐPQÑ0QÑQÐTXÔTbÑbñ!
ô !
Ðõ œ) DÔ$4Ô$6Ð8MÑNÔNˆõ !œfÝŒG�O¨4Ô+DÈDÐQÑQÔQÐTXÔTkÑkñ
ô 
Ðð
 #Ô$Ð$rO   Únoise_lengthÚ	generatorc                 óš   — |€t           j                             ¦   «         }|| j        f}|                     |t           j        ¬¦  «        }|S )añ  
        Generates a random noise sequence of standard Gaussian noise for use in the `noise_sequence` argument of
        [`UnivNetModel.forward`].

        Args:
            spectrogram_length (`int`):
                The length (dim 0) of the generated noise.
            model_in_channels (`int`, *optional*, defaults to `None`):
                The number of features (dim 1) of the generated noise. This should correspond to the
                `model_in_channels` of the [`UnivNetGan`] model. If not set, this will default to
                `self.config.model_in_channels`.
            generator (`numpy.random.Generator`, *optional*, defaults to `None`)
                An optional `numpy.random.Generator` random number generator to control noise generation. If not set, a
                new generator with fresh entropy will be created.

        Returns:
            `numpy.ndarray`: Array containing random standard Gaussian noise of shape `(noise_length,
            model_in_channels)`.
        N©Údtype)r`   ÚrandomÚdefault_rngr3   Ústandard_normalÚfloat32)rK   rn   ro   Únoise_shapeÚnoises        rN   Úgenerate_noisez&UnivNetFeatureExtractor.generate_noiseç   sI   € ð0 ÐÝœ	×-Ò-Ñ/Ô/ˆIà# TÔ%;Ð<ˆØ×)Ò)¨+½R¼ZÐ)ÑHÔHˆàˆrO   c                 óZ   ‡— d„ |D ¦   «         }‰�ˆfd„t          |¦  «        D ¦   «         }|S )aÏ  
        Removes padding from generated audio after running [`UnivNetModel.forward`]. This returns a ragged list of 1D
        audio waveform arrays and not a single tensor/array because in general the waveforms will have different
        lengths after removing padding.

        Args:
            waveforms (`torch.FloatTensor` of shape `(batch_size, sequence_length)`):
                The batched output waveforms from the [`UnivNetModel`].
            waveform_lengths (`torch.FloatTensor` of shape `(batch_size,)`, *optional*):
                The batched lengths of each waveform before padding.

        Returns:
            `list[np.ndarray]`: A ragged list of 1D waveform arrays with padding removed.
        c                 ó„   — g | ]=}|                      ¦   «                              d d¬¦  «                             ¦   «         ‘Œ>S )ÚcpuT)ÚdeviceÚcopy)ÚdetachÚtoÚnumpy)Ú.0rV   s     rN   ú
<listcomp>z8UnivNetFeatureExtractor.batch_decode.<locals>.<listcomp>  sA   € ÐeÐeÐeÈx�X—_’_Ñ&Ô&×)Ò)°¸TÐ)ÑBÔB×HÒHÑJÔJÐeÐeÐerO   Nc                 ó6   •— g | ]\  }}|d ‰|         …         ‘ŒS ©NrB   )r‚   ÚirV   Úwaveform_lengthss      €rN   rƒ   z8UnivNetFeatureExtractor.batch_decode.<locals>.<listcomp>  s.   ø€ ÐaÐaÐa¹[¸QÀ˜Ð"7Ð$4°QÔ$7Ð"7Ô8ÐaÐaÐarO   )Ú	enumerate)rK   Ú	waveformsr‡   s     `rN   Úbatch_decodez$UnivNetFeatureExtractor.batch_decode  sH   ø€ ð  fÐeÐ[dÐeÑeÔeˆ	àÐ'ØaÐaÐaÐaÍIÐV_ÑL`ÔL`ÐaÑaÔaˆIàÐrO   Ú
raw_speechÚpaddingÚ
max_lengthÚ
truncationÚpad_to_multiple_ofÚreturn_noiseÚpad_endÚ
pad_lengthr6   Úreturn_tensorsc                 ó  ‡ ‡‡
— |�|n‰ j         }|�<|‰ j        k    r0t          d‰ j        j        › d‰ j        › d‰ j        › d|› d�	¦  «        ‚n(t
                               d‰ j        j        › d�¦  «         t          |t          j	        ¦  «        ot          |j        ¦  «        d	k    }|r*t          |j        ¦  «        d
k    rt          d‰ › �¦  «        ‚|pHt          |t          t          f¦  «        o,t          |d         t          j	        t          t          f¦  «        }|rd„ |D ¦   «         }n›|s;t          |t          j	        ¦  «        s!t          j        |t          j        ¬¦  «        }n^t          |t          j	        ¦  «        rD|j        t          j        t          j        ¦  «        u r|                     t          j        ¦  «        }|s!t          j        |t          j        ¬¦  «        g}|	r‰
�‰
n‰ j        Š
ˆ
ˆ fd„|D ¦   «         }t)          d|i¦  «        }‰                      |||�|n‰ j        |||¬¦  «        }|                     d¦  «        }ˆ fd„|D ¦   «         }t          |d         t          ¦  «        rd„ |D ¦   «         |d<   nd„ |D ¦   «         |d<   |                     d¦  «        }|�d„ |D ¦   «         |d<   |rˆˆ fd„|d         D ¦   «         }||d<   |rˆ fd„|d         D ¦   «         |d<   |�|                     |¦  «        }|S )a„  
        Main method to featurize and prepare for the model one or several sequence(s).

        Args:
            raw_speech (`np.ndarray`, `list[float]`, `list[np.ndarray]`, `list[list[float]]`):
                The sequence or batch of sequences to be padded. Each sequence can be a numpy array, a list of float
                values, a list of numpy arrays or a list of list of float values. Must be mono channel audio, not
                stereo, i.e. single float per timestep.
            sampling_rate (`int`, *optional*):
                The sampling rate at which the `raw_speech` input was sampled. It is strongly recommended to pass
                `sampling_rate` at the forward call to prevent silent errors and allow automatic speech recognition
                pipeline.
            padding (`bool`, `str` or [`~utils.PaddingStrategy`], *optional*, defaults to `True`):
                Select a strategy to pad the input `raw_speech` waveforms (according to the model's padding side and
                padding index) among:

                - `True` or `'longest'`: Pad to the longest sequence in the batch (or no padding if only a single
                  sequence if provided).
                - `'max_length'`: Pad to a maximum length specified with the argument `max_length` or to the maximum
                  acceptable input length for the model if that argument is not provided.
                - `False` or `'do_not_pad'` (default): No padding (i.e., can output a batch with sequences of different
                  lengths).

                If `pad_end = True`, that padding will occur before the `padding` strategy is applied.
            max_length (`int`, *optional*):
                Maximum length of the returned list and optionally padding length (see above).
            truncation (`bool`, *optional*, defaults to `True`):
                Activates truncation to cut input sequences longer than `max_length` to `max_length`.
            pad_to_multiple_of (`int`, *optional*):
                If set will pad the sequence to a multiple of the provided value.

                This is especially useful to enable the use of Tensor Cores on NVIDIA hardware with compute capability
                `>= 7.5` (Volta), or on TPUs which benefit from having sequence lengths be a multiple of 128.
            return_noise (`bool`, *optional*, defaults to `True`):
                Whether to generate and return a noise waveform for use in [`UnivNetModel.forward`].
            generator (`numpy.random.Generator`, *optional*, defaults to `None`):
                An optional `numpy.random.Generator` random number generator to use when generating noise.
            pad_end (`bool`, *optional*, defaults to `False`):
                Whether to pad the end of each waveform with silence. This can help reduce artifacts at the end of the
                generated audio sample; see https://github.com/seungwonpark/melgan/issues/8 for more details. This
                padding will be done before the padding strategy specified in `padding` is performed.
            pad_length (`int`, *optional*, defaults to `None`):
                If padding the end of each waveform, the length of the padding in spectrogram frames. If not set, this
                will default to `self.config.pad_end_length`.
            do_normalize (`bool`, *optional*):
                Whether to perform Tacotron 2 normalization on the input. Normalizing can help to significantly improve
                the performance for some models. If not set, this will default to `self.config.do_normalize`.
            return_attention_mask (`bool`, *optional*):
                Whether to return the attention mask. If left to the default, will return the attention mask according
                to the specific feature_extractor's default.

                [What are attention masks?](../glossary#attention-mask)

            return_tensors (`str` or [`~utils.TensorType`], *optional*):
                If set, will return tensors instead of list of python integers. Acceptable values are:

                - `'pt'`: Return PyTorch `torch.np.array` objects.
                - `'np'`: Return Numpy `np.ndarray` objects.
        Nz3The model corresponding to this feature extractor: z& was trained using a sampling rate of zI. Please make sure that the provided `raw_speech` input was sampled with z	 and not ú.zDIt is strongly recommended to pass the `sampling_rate` argument to `zN()`. Failing to do so can result in silent errors that might be hard to debug.r   r7   z2Only mono-channel audio is supported for input to r   c                 óN   — g | ]"}t          j        |t           j        ¬ ¦  «        ‘Œ#S ©rq   ©r`   Úasarrayrv   )r‚   Úspeechs     rN   rƒ   z4UnivNetFeatureExtractor.__call__.<locals>.<listcomp>€  s)   € ÐXÐXÐXÀ6�"œ* Vµ2´:Ð>Ñ>Ô>ÐXÐXÐXrO   rq   c                 ó\   •— g | ](}t          j        |d ‰‰j        z  f‰j        ¬¦  «        ‘Œ)S )r   )Úconstant_values)r`   ra   r&   r#   )r‚   rV   r’   rK   s     €€rN   rƒ   z4UnivNetFeatureExtractor.__call__.<locals>.<listcomp>�  sJ   ø€ ð ð ð àõ ”�x ! Z°$´/Ñ%AÐ!BÐTXÔTfÐgÑgÔgðð ð rO   r   )rŒ   r�   rŽ   r�   r6   c                 ó:   •— g | ]}‰                      |¦  «        ‘ŒS rB   )rl   )r‚   rV   rK   s     €rN   rƒ   z4UnivNetFeatureExtractor.__call__.<locals>.<listcomp>¡  s'   ø€ ÐZÐZÐZ¸x˜D×0Ò0°Ñ:Ô:ÐZÐZÐZrO   c                 óN   — g | ]"}t          j        |t           j        ¬ ¦  «        ‘Œ#S r—   r˜   ©r‚   Úmels     rN   rƒ   z4UnivNetFeatureExtractor.__call__.<locals>.<listcomp>¤  s*   € Ð/nÐ/nÐ/nÐVYµ´
¸3ÅbÄjÐ0QÑ0QÔ0QÐ/nÐ/nÐ/nrO   c                 óL   — g | ]!}|                      t          j        ¦  «        ‘Œ"S rB   )Úastyper`   rv   rŸ   s     rN   rƒ   z4UnivNetFeatureExtractor.__call__.<locals>.<listcomp>¦  s&   € Ð/cÐ/cÐ/cÈ3°·
²
½2¼:Ñ0FÔ0FÐ/cÐ/cÐ/crO   Úattention_maskc                 óN   — g | ]"}t          j        |t           j        ¬ ¦  «        ‘Œ#S r—   )r`   r™   Úint32)r‚   Úarrays     rN   rƒ   z4UnivNetFeatureExtractor.__call__.<locals>.<listcomp>«  s*   € Ð-lÐ-lÐ-lÐTY­b¬j¸ÅbÄhÐ.OÑ.OÔ.OÐ-lÐ-lÐ-lrO   r   c                 óR   •— g | ]#}‰                      |j        d          ‰¦  «        ‘Œ$S )r   )ry   Úshape)r‚   r   ro   rK   s     €€rN   rƒ   z4UnivNetFeatureExtractor.__call__.<locals>.<listcomp>®  sA   ø€ ð ð ð àð ×#Ò# KÔ$5°aÔ$8¸)ÑDÔDðð ð rO   r   c                 ó:   •— g | ]}‰                      |¦  «        ‘ŒS rB   )rS   )r‚   r   rK   s     €rN   rƒ   z4UnivNetFeatureExtractor.__call__.<locals>.<listcomp>µ  s2   ø€ ð 0ð 0ð 0Ø0;�—’˜{Ñ+Ô+ð0ð 0ð 0rO   )r$   r"   Ú
ValueErrorrM   Ú__name__ÚloggerÚwarningÚ
isinstancer`   ÚndarrayÚlenr¨   ÚlistÚtupler™   rv   rr   Úfloat64r¢   r4   r
   ra   rF   ÚgetÚconvert_to_tensors)rK   r‹   r"   rŒ   r�   rŽ   r�   r�   ro   r‘   r’   r$   r6   r“   Úis_batched_numpyÚ
is_batchedÚbatched_speechÚpadded_inputsr   Úmel_spectrogramsr£   rx   s   `       ` `           rN   Ú__call__z UnivNetFeatureExtractor.__call__  s  øøø€ ðV (4Ð'?�|�|ÀTÔEVˆàÐ$Ø Ô 2Ò2Ð2Ý ðWÈ$Ì.ÔJað Wð WØ)-Ô);ðWð Wà)-Ô);ðWð WàFSðWð Wð Wñô ð ð 3õ �NŠNð\ÐW[ÔWeÔWnð \ð \ð \ñô ð õ
 & jµ"´*Ñ=Ô=Ð[Å#ÀjÔFVÑBWÔBWÐZ[ÒB[ÐØð 	Z¥ JÔ$4Ñ 5Ô 5¸Ò 9Ð 9ÝÐXÐRVÐXÐXÑYÔYÐYØ%ð 
Ý�z¥D­% =Ñ1Ô1ÐlµzÀ*ÈQÄ-ÕRTÔR\Õ^cÕeiÐQjÑ7kÔ7kð 	ð ð 	7ØXÐXÈZÐXÑXÔXˆJˆJØð 	7¥J¨z½2¼:Ñ$FÔ$Fð 	7Ýœ Jµb´jÐAÑAÔAˆJˆJÝ˜
¥B¤JÑ/Ô/ð 	7°JÔ4DÍÌÕQSÔQ[ÑH\ÔH\Ð4\Ð4\Ø#×*Ò*­2¬:Ñ6Ô6ˆJð ð 	DÝœ* Zµr´zÐBÑBÔBÐCˆJð ð 	Ø'1Ð'=˜˜À4ÔCVˆJðð ð ð ð à *ðñ ô ˆJõ
 &Ð'7¸Ð&DÑEÔEˆàŸšØØØ%/Ð%;�z�zÀÔAUØ!Ø1Ø"7ð !ñ 
ô 
ˆð '×*Ò*Ð+;Ñ<Ô<ˆàZÐZÐZÐZÈ>ÐZÑZÔZÐå�n QÔ'­Ñ.Ô.ð 	dØ/nÐ/nÐ]mÐ/nÑ/nÔ/nˆNÐ+Ñ,Ð,à/cÐ/cÐRbÐ/cÑ/cÔ/cˆNÐ+Ñ,ð '×*Ò*Ð+;Ñ<Ô<ˆØÐ%Ø-lÐ-lÐ]kÐ-lÑ-lÔ-lˆN˜>Ñ*àð 	5ðð ð ð ð à#1Ð2BÔ#Cðñ ô ˆEð 05ˆNÐ+Ñ,àð 	ð0ð 0ð 0ð 0Ø?MÐN^Ô?_ð0ñ 0ô 0ˆNÐ+Ñ,ð Ð%Ø+×>Ò>¸~ÑNÔNˆNàÐrO   c                 óh   •— t          ¦   «                              ¦   «         }g d¢}|D ]	}||v r||= Œ
|S )N)rI   rJ   rG   rH   rF   )rC   Úto_dict)rK   ÚoutputÚnamesr9   rM   s       €rN   r½   zUnivNetFeatureExtractor.to_dict¾  sI   ø€ Ý‘”—’Ñ"Ô"ˆð QÐPÐPˆØð 	!ð 	!ˆDØ�vˆ~ˆ~Ø˜4�LøàˆrO   )r   r   r   Fr   r   r   r   r   r   r   Nr   Fr   r   r   r   r    r   Tr…   )NTNTNTNFNNNN)r«   Ú
__module__Ú__qualname__Ú__doc__Úmodel_input_namesrb   rE   ÚboolÚstrrD   rS   rU   r`   r¯   rl   rs   Ú	Generatorry   r±   rŠ   r   r   r
   r»   Údictr   r½   Ú__classcell__)rM   s   @rN   r   r      s¦  ø€ € € € € ðCð CðJ MÐLÐLÐð Ø"Ø"Ø"ØØØØ)Ø$(ØØØ!ØØØ$'Ø&*Ø2Ø1Ø!#Ø Ø"ð-J-ð J-àðJ-ð ðJ-ð ð	J-ð
 ðJ-ð ðJ-ð ðJ-ð ðJ-ð ðJ-ð ˜T‘zðJ-ð ðJ-ð ðJ-ð �d‰lðJ-ð ðJ-ð ðJ-ð  "ð!J-ð" $ð#J-ð$ ð%J-ð& ð'J-ð( ð)J-ð* ð+J-ð J-ð J-ð J-ð J-ð J-ðXhð hð hðhð hð hð.%¨¬
ð .%°r´zð .%ð .%ð .%ð .%ðf 15ðð àðð ”9Ô&¨Ñ-ðð 
Œð	ð ð ð ð@ð ÀÀRÄZÔ@Pð ð ð ð ð4 %)Ø04Ø!%ØØ)-Ø!Ø04ØØ!%Ø#'Ø-1Ø26ð^ð ^à”J  e¤Ñ,¨t°B´JÔ/?Ñ?À$ÀtÈEÄ{ÔBSÑSð^ð ˜T‘zð^ð ˜‘˜oÑ-ð	^ð
 ˜$‘Jð^ð ð^ð   $™Jð^ð ð^ð ”9Ô&¨Ñ-ð^ð ð^ð ˜$‘Jð^ð ˜D‘jð^ð  $ d™{ð^ð ˜jÑ(¨4Ñ/ð^ð 
ð^ð ^ð ^ð ^ð@	˜˜c 3˜hœð 	ð 	ð 	ð 	ð 	ð 	ð 	ð 	ð 	ð 	rO   r   )rÂ   Útypingr   r�   r`   Úaudio_utilsr   r   r   r   Ú!feature_extraction_sequence_utilsr	   Úfeature_extraction_utilsr
   Úutilsr   r   r   Ú
get_loggerr«   r¬   r   Ú__all__rB   rO   rN   ú<module>rÐ      sç   ðð 0Ð /à Ð Ð Ð Ð Ð à Ð Ð Ð à \Ð \Ð \Ð \Ð \Ð \Ð \Ð \Ð \Ð \Ð \Ð \Ø IÐ IÐ IÐ IÐ IÐ IØ 4Ð 4Ð 4Ð 4Ð 4Ð 4Ø 9Ð 9Ð 9Ð 9Ð 9Ð 9Ð 9Ð 9Ð 9Ð 9ð 
ˆÔ	˜HÑ	%Ô	%€ðjð jð jð jð jÐ6ñ jô jð jðZ %Ð
%€€€rO   