§
    9²¸j±¾  ã                   óð  — d dl Z d dlmZ d dlmZ d dlZd dlZd dlm	Z	 d dl
m	c mZ d dlmZ ddlmZ ddlmZ ddlmZ dd	lmZmZ dd
lmZ ddlmZmZ ddlmZmZ ddl m!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, ddl-m.Z. e%e G d„ de#¦  «        ¦   «         ¦   «         Z/e%e G d„ de#¦  «        ¦   «         ¦   «         Z0e%e G d„ de#¦  «        ¦   «         ¦   «         Z1 G d„ de	j2        ¦  «        Z3 G d„ de	j2        ¦  «        Z4d„ Z5 ed ¦  «        dXd!„¦   «         Z6d"ej7        d#e8d$ej7        fd%„Z9	 dYd'e	j2        d(ej7        d)ej7        d*ej7        d+ej7        dz  d,e:d-e:d.e!e$         fd/„Z; ee6¦  «         G d0„ d1e	j2        ¦  «        ¦   «         Z< ed2¦  «         G d3„ d4e	j2        ¦  «        ¦   «         Z= G d5„ d6e¦  «        Z> G d7„ d8e	j2        ¦  «        Z?d9„ Z@ G d:„ d;e	j2        ¦  «        ZA G d<„ d=e	j2        ¦  «        ZB G d>„ d?e	j2        ¦  «        ZC G d@„ dAe	j2        ¦  «        ZD G dB„ dCe	j2        ¦  «        ZE G dD„ dEe	j2        ¦  «        ZF G dF„ dGe	j2        ¦  «        ZG G dH„ dIe	j2        ¦  «        ZH G dJ„ dKe	j2        ¦  «        ZI G dL„ dMe	j2        ¦  «        ZJ G dN„ dOe	j2        ¦  «        ZK G dP„ dQe	j2        ¦  «        ZLe% G dR„ dSe¦  «        ¦   «         ZM e%dT¬U¦  «         G dV„ dWeM¦  «        ¦   «         ZNdWdSgZOdS )Zé    N)ÚCallable)Ú	dataclass)Ú	Parameteré   )Úinitialization)ÚACT2FN)ÚCache)Úuse_kernel_forward_from_hubÚuse_kernelized_func)ÚGradientCheckpointingLayer)ÚROPE_INIT_FUNCTIONSÚdynamic_rope_update)ÚALL_ATTENTION_FUNCTIONSÚPreTrainedModel)ÚUnpack)ÚModelOutputÚTransformersKwargsÚauto_docstringÚcan_return_tuple)Údeprecate_kwarg)Úmaybe_autocasté   )Ú	AutoModelé   )ÚXcodec2Configc                   óŒ   — e Zd ZU dZdZej        dz  ed<   dZej	        dz  ed<   dZ
ej        dz  ed<   dZej        dz  ed<   dS )ÚXcodec2OutputaL  
    audio_values (`torch.FloatTensor` of shape `(batch_size, 1, sequence_length)`, *optional*):
        Decoded audio waveform values in the time domain, obtained using the decoder
        part of Xcodec2. These represent the reconstructed audio signal.
    audio_codes (`torch.LongTensor` of shape `(batch_size, 1, codes_length)`, *optional*):
        Discrete code embeddings computed using `model.encode`. These are the quantized
        representations of the input audio used for further processing or generation.
    latents (`torch.Tensor` of shape `(batch_size, dimension, time_steps)`):
        Quantized continuous representation of input's embedding.
    audio_codes_mask (`torch.int32` of shape `(batch_size, 1, codes_length)`, *optional*):
        Downsampled `padding_mask` for indicating valid audio codes in `audio_codes`.
    NÚaudio_valuesÚaudio_codesÚlatentsÚaudio_codes_mask)Ú__name__Ú
__module__Ú__qualname__Ú__doc__r   ÚtorchÚFloatTensorÚ__annotations__r   Ú
LongTensorr    ÚTensorr!   © ó    ú|/root/private/project/pallet_detection_v2/.venv/lib/python3.11/site-packages/transformers/models/xcodec2/modeling_xcodec2.pyr   r   -   s}   € € € € € € ðð ð .2€L�%Ô# dÑ*Ð1Ð1Ñ1Ø+/€K�Ô! DÑ(Ð/Ð/Ñ/Ø#'€GˆUŒ\˜DÑ Ð'Ð'Ñ'Ø,0Ð�e”l TÑ)Ð0Ð0Ñ0Ð0Ð0r,   r   c                   ón   — e Zd ZU dZdZej        dz  ed<   dZej	        dz  ed<   dZ
ej	        dz  ed<   dS )ÚXcodec2EncoderOutputat  
    audio_codes (`torch.LongTensor` of shape `(batch_size, 1, codes_length)`, *optional*):
        Discrete code embeddings computed using `model.encode`. These represent
        the compressed, quantized form of the input audio signal that can be
        used for storage, transmission, or generation.
    latents (`torch.Tensor` of shape `(batch_size, dimension, time_steps)`):
        Quantized continuous representation of input's embedding.
    audio_codes_mask (`torch.int32` of shape `(batch_size, 1, codes_length)`, *optional*):
        Downsampled `padding_mask` for indicating valid audio codes in `audio_codes`.
    Nr   r    r!   )r"   r#   r$   r%   r   r&   r)   r(   r    r*   r!   r+   r,   r-   r/   r/   C   se   € € € € € € ð	ð 	ð ,0€K�Ô! DÑ(Ð/Ð/Ñ/Ø#'€GˆUŒ\˜DÑ Ð'Ð'Ñ'Ø,0Ð�e”l TÑ)Ð0Ð0Ñ0Ð0Ð0r,   r/   c                   ó2   — e Zd ZU dZdZej        dz  ed<   dS )ÚXcodec2DecoderOutputa=  
    audio_values (`torch.FloatTensor` of shape `(batch_size, 1, segment_length)`, *optional*):
        Decoded audio waveform values in the time domain, obtained by converting
        the discrete codes back into continuous audio signals. This represents
        the reconstructed audio that can be played back.
    Nr   )r"   r#   r$   r%   r   r&   r'   r(   r+   r,   r-   r1   r1   V   s6   € € € € € € ðð ð .2€L�%Ô# dÑ*Ð1Ð1Ñ1Ð1Ð1r,   r1   c            	       óî   ‡ — e Zd Z edd¬¦  «        d
defˆ fd„¦   «         Ze edd¬¦  «        d
dedeej	        e
f         fd„¦   «         ¦   «         Z ej        ¦   «         ed	„ ¦   «         ¦   «         Zˆ xZS )ÚXcodec2RotaryEmbeddingÚdevicez5.18)ÚversionNÚconfigc                 ó¾  •— t          ¦   «                              ¦   «          |j        | _        |j        | _        || _        | j        j        d         | _        | j        }| j        dk    rt          | j                 } || j        |¦  «        \  }| _
        t          j        |d¬¦  «        | _        t          j        |                     ¦   «         d¬¦  «        | _        d S )NÚ	rope_typeÚdefaultF©Ú
persistent)ÚsuperÚ__init__Úmax_position_embeddingsÚmax_seq_len_cachedÚoriginal_max_seq_lenr6   Úrope_parametersr8   Úcompute_default_rope_parametersr   Úattention_scalingÚnnÚBufferÚinv_freqÚcloneÚoriginal_inv_freq)Úselfr6   r4   Úrope_init_fnrF   Ú	__class__s        €r-   r=   zXcodec2RotaryEmbedding.__init__d   sÀ   ø€ å‰Œ×ÒÑÔÐØ"(Ô"@ˆÔØ$*Ô$BˆÔ!àˆŒàœÔ4°[ÔAˆŒØ!%Ô!EˆØŒ>˜YÒ&Ð&Ý.¨t¬~Ô>ˆLØ+7¨<¸¼ÀVÑ+LÔ+LÑ(ˆ�$Ô(åœ	 (°uÐ=Ñ=Ô=ˆŒÝ!#¤¨8¯>ª>Ñ+;Ô+;ÈÐ!NÑ!NÔ!NˆÔÐÐr,   Úreturnc                 óä   — | j         d         }t          | dd¦  «        p| j        | j        z  }d}d|t	          j        d|dt          j        ¬¦  «        |z  z  z  }|                     |¦  «        |fS )a¾  
        Computes the inverse frequencies according to the original RoPE implementation
        Args:
            config ([`~transformers.PreTrainedConfig`]):
                The model configuration.
        Returns:
            Tuple of (`torch.Tensor`, `float`), containing the inverse frequencies for the RoPE embeddings and the
            post-processing scaling factor applied to the computed cos/sin (unused in this type of RoPE).
        Ú
rope_thetaÚhead_dimNç      ð?r   r   ©Údtype)rA   ÚgetattrÚhidden_sizeÚnum_attention_headsr&   ÚarangeÚfloatÚto)r6   r4   ÚkwargsÚbaseÚdimÚattention_factorrF   s          r-   rB   z6Xcodec2RotaryEmbedding.compute_default_rope_parametersu   sz   € ð Ô% lÔ3ˆÝ�f˜j¨$Ñ/Ô/Ðc°6Ô3EÈÔIcÑ3cˆàÐà˜$¥5¤<°°3¸ÅÄÐ#MÑ#MÔ#MÐPSÑ#SÑTÑUˆØ�{Š{˜6Ñ"Ô"Ð$4Ð4Ð4r,   c                 óú  — | j         d d d …d f                              |j        d         dd¦  «                             t          j        |j        ¬¦  «        }|d d …d d d …f                              ¦   «         }t          |j        j        t          ¦  «        r|j        j        dk    r|j        j        nd}t          |d¬¦  «        5  ||z                       dd	¦  «        }t	          j        ||fd¬
¦  «        }|                     ¦   «         | j        z  }|                     ¦   «         | j        z  }	d d d ¦  «         n# 1 swxY w Y   |                     |j        ¬¦  «        |	                     |j        ¬¦  «        fS )Nr   éÿÿÿÿr   ©rR   r4   ÚmpsÚcpuF©Údevice_typeÚenabledr   ©r[   rQ   )rF   ÚexpandÚshaperX   r&   rW   r4   Ú
isinstanceÚtypeÚstrr   Ú	transposeÚcatÚcosrC   ÚsinrR   )
rI   ÚxÚposition_idsÚinv_freq_expandedÚposition_ids_expandedrc   ÚfreqsÚembrm   rn   s
             r-   ÚforwardzXcodec2RotaryEmbedding.forward‰   s›  € ð ŒM˜$    4˜-Ô(×/Ò/°Ô0BÀ1Ô0EÀrÈ1ÑMÔM×PÒPÕW\ÔWbÐklÔksÐPÑtÔtð 	ð !-¨Q¨Q¨Q°°a°a°a¨ZÔ 8× >Ò >Ñ @Ô @Ðå'1°!´(´-ÅÑ'EÔ'EÐkÈ!Ì(Ì-Ð[`ÒJ`ÐJ`�a”h”m�mÐfkˆå¨¸UÐCÑCÔCð 	5ð 	5Ø&Ð)>Ñ>×IÒIÈ!ÈQÑOÔOˆEÝ”)˜U E˜N°Ð3Ñ3Ô3ˆCØ—'’'‘)”)˜dÔ4Ñ4ˆCØ—'’'‘)”)˜dÔ4Ñ4ˆCð		5ð 	5ð 	5ñ 	5ô 	5ð 	5ð 	5ð 	5ð 	5ð 	5ð 	5øøøð 	5ð 	5ð 	5ð 	5ð �vŠv˜AœGˆvÑ$Ô$ c§f¢f°1´7 fÑ&;Ô&;Ð;Ð;s   ÃA*D<Ä<E ÅE ©N)r"   r#   r$   r   r   r=   ÚstaticmethodÚtupler&   r*   rW   rB   Úno_gradr   ru   Ú__classcell__©rK   s   @r-   r3   r3   c   sö   ø€ € € € € Ø€_�X vÐ.Ñ.Ô.ðOð O˜}ð Oð Oð Oð Oð Oñ /Ô.ðOð  Ø€_�X vÐ.Ñ.Ô.ð5ð 5°ð 5ÐY^Ð_dÔ_kÐmrÐ_rÔYsð 5ð 5ð 5ñ /Ô.ñ „\ð5ð$ €U„]�_„_Øð<ð <ñ Ôñ „_ð<ð <ð <ð <ð <r,   r3   c                   óH   ‡ — e Zd Zdefˆ fd„Zdej        dej        fd„Zˆ xZS )Ú
Xcodec2MLPr6   c                 ó  •— t          ¦   «                              ¦   «          || _        t          |j                 | _        t          j        |j        |j	        d¬¦  «        | _
        t          j        |j	        |j        d¬¦  «        | _        d S )NF©Úbias)r<   r=   r6   r   Ú
hidden_actÚactivation_fnrD   ÚLinearrT   Úintermediate_sizeÚfc1Úfc2©rI   r6   rK   s     €r-   r=   zXcodec2MLP.__init__�   sr   ø€ Ý‰Œ×ÒÑÔÐØˆŒÝ# FÔ$5Ô6ˆÔÝ”9˜VÔ/°Ô1IÐPUÐVÑVÔVˆŒÝ”9˜VÔ5°vÔ7IÐPUÐVÑVÔVˆŒˆˆr,   Úhidden_statesrL   c                 ó„   — |                       |¦  «        }|                      |¦  «        }|                      |¦  «        }|S rv   )r…   r‚   r†   ©rI   rˆ   s     r-   ru   zXcodec2MLP.forward¤   s=   € ØŸš Ñ/Ô/ˆØ×*Ò*¨=Ñ9Ô9ˆØŸš Ñ/Ô/ˆØÐr,   ©	r"   r#   r$   r   r=   r&   r*   ru   rz   r{   s   @r-   r}   r}   œ   sq   ø€ € € € € ðW˜}ð Wð Wð Wð Wð Wð Wð U¤\ð °e´lð ð ð ð ð ð ð ð r,   r}   c                 óœ   — | dd| j         d         dz  …f         }| d| j         d         dz  d…f         }t          j        | |fd¬¦  «        S )z*Rotates half the hidden dims of the input..Nr^   r   re   )rg   r&   rl   )ro   Úx1Úx2s      r-   Úrotate_halfr�   «   s]   € à	
ˆ3Ð"�!”'˜"”+ Ñ"Ð"Ð"Ô	#€BØ	
ˆ3�”˜”˜qÑ Ð"Ð"Ð"Ô	#€BÝŒ9�r�c˜2�Y BÐ'Ñ'Ô'Ð'r,   Úrotary_pos_embc                 ó¾   — |                      |¦  «        }|                      |¦  «        }| |z  t          | ¦  «        |z  z   }||z  t          |¦  «        |z  z   }||fS )a…  Applies Rotary Position Embedding to the query and key tensors.

    Args:
        q (`torch.Tensor`): The query tensor.
        k (`torch.Tensor`): The key tensor.
        cos (`torch.Tensor`): The cosine part of the rotary embedding.
        sin (`torch.Tensor`): The sine part of the rotary embedding.
        unsqueeze_dim (`int`, *optional*, defaults to 1):
            The 'unsqueeze_dim' argument specifies the dimension along which to unsqueeze cos[position_ids] and
            sin[position_ids] so that they can be properly broadcasted to the dimensions of q and k. For example, note
            that cos[position_ids] and sin[position_ids] have the shape [batch_size, seq_len, head_dim]. Then, if q and
            k have the shape [batch_size, heads, seq_len, head_dim], then setting unsqueeze_dim=1 makes
            cos[position_ids] and sin[position_ids] broadcastable to the shapes of q and k. Similarly, if q and k have
            the shape [batch_size, seq_len, heads, head_dim], then set unsqueeze_dim=2.
    Returns:
        `tuple(torch.Tensor)` comprising of the query and key tensors rotated using the Rotary Position Embedding.
    )Ú	unsqueezer�   )ÚqÚkrm   rn   Úunsqueeze_dimÚq_embedÚk_embeds          r-   Úapply_rotary_pos_embr˜   ²   sc   € ð& �-Š-˜Ñ
&Ô
&€CØ
�-Š-˜Ñ
&Ô
&€CØ�3‰w�; q™>œ>¨CÑ/Ñ0€GØ�3‰w�; q™>œ>¨CÑ/Ñ0€GØ�GÐÐr,   rˆ   Ún_reprL   c                 ó¸   — | j         \  }}}}|dk    r| S | dd…dd…ddd…dd…f                              |||||¦  «        } |                      |||z  ||¦  «        S )zÔ
    This is the equivalent of torch.repeat_interleave(x, dim=1, repeats=n_rep). The hidden states go from (batch,
    num_key_value_heads, seqlen, head_dim) to (batch, num_attention_heads, seqlen, head_dim)
    r   N)rg   rf   Úreshape)rˆ   r™   ÚbatchÚnum_key_value_headsÚslenrO   s         r-   Ú	repeat_kvrŸ   Ì   s„   € ð
 2?Ô1DÑ.€EÐ  hØ�‚z€zØÐØ! ! ! ! Q Q Q¨¨a¨a¨a°°°Ð"2Ô3×:Ò:¸5ÐBUÐW\Ð^bÐdlÑmÔm€MØ× Ò  Ð(;¸eÑ(CÀTÈ8ÑTÔTÐTr,   ç        ÚmoduleÚqueryÚkeyÚvalueÚattention_maskÚscalingÚdropoutrY   c                 ó  — t          || j        ¦  «        }t          || j        ¦  «        }	t          j        ||                     dd¦  «        ¦  «        |z  }
|�|
|z   }
t
          j                             |
dt          j        ¬¦  «         	                    |j
        ¦  «        }
t
          j                             |
|| j        ¬¦  «        }
t          j        |
|	¦  «        }|                     dd¦  «                             ¦   «         }||
fS )Nr   r   r^   ©r[   rR   ©ÚpÚtrainingr   )rŸ   Únum_key_value_groupsr&   Úmatmulrk   rD   Ú
functionalÚsoftmaxÚfloat32rX   rR   r§   r¬   Ú
contiguous)r¡   r¢   r£   r¤   r¥   r¦   r§   rY   Ú
key_statesÚvalue_statesÚattn_weightsÚattn_outputs               r-   Úeager_attention_forwardr·   Ø   sé   € õ ˜3 Ô ;Ñ<Ô<€JÝ˜U FÔ$?Ñ@Ô@€Lå”<  z×';Ò';¸A¸qÑ'AÔ'AÑBÔBÀWÑL€LØÐ!Ø# nÑ4ˆå”=×(Ò(¨¸2ÅUÄ]Ð(ÑSÔS×VÒVÐW\ÔWbÑcÔc€LÝ”=×(Ò(¨¸È6Ì?Ð(Ñ[Ô[€LÝ”,˜|¨\Ñ:Ô:€KØ×'Ò'¨¨1Ñ-Ô-×8Ò8Ñ:Ô:€Kà˜Ð$Ð$r,   c                   óÎ   ‡ — e Zd ZdZdedefˆ fd„Z	 	 	 ddej        de	ej        ej        f         dz  dej        dz  d	e
dz  d
ee         de	ej        ej        f         fd„Zˆ xZS )ÚXcodec2Attentionz=Multi-headed attention from 'Attention Is All You Need' paperr6   Ú	layer_idxc                 ó®  •— t          ¦   «                              ¦   «          || _        || _        t	          |d|j        |j        z  ¦  «        | _        |j        |j        z  | _	        | j        dz  | _
        |j        | _        d| _        t          j        |j        |j        | j        z  |j        ¬¦  «        | _        t          j        |j        |j        | j        z  |j        ¬¦  «        | _        t          j        |j        |j        | j        z  |j        ¬¦  «        | _        t          j        |j        | j        z  |j        |j        ¬¦  «        | _        d S )NrO   g      à¿Fr   )r<   r=   r6   rº   rS   rT   rU   rO   r�   r­   r¦   Úattention_dropoutÚ	is_causalrD   rƒ   Úattention_biasÚq_projÚk_projÚv_projÚo_proj©rI   r6   rº   rK   s      €r-   r=   zXcodec2Attention.__init__õ   sB  ø€ Ý‰Œ×ÒÑÔÐØˆŒØ"ˆŒÝ ¨
°FÔ4FÈ&ÔJdÑ4dÑeÔeˆŒØ$*Ô$>À&ÔB\Ñ$\ˆÔ!Ø”} dÑ*ˆŒØ!'Ô!9ˆÔØˆŒå”iØÔ Ô :¸T¼]Ñ JÐQWÔQfð
ñ 
ô 
ˆŒõ ”iØÔ Ô :¸T¼]Ñ JÐQWÔQfð
ñ 
ô 
ˆŒõ ”iØÔ Ô :¸T¼]Ñ JÐQWÔQfð
ñ 
ô 
ˆŒõ ”iØÔ&¨¬Ñ6¸Ô8JÐQWÔQfð
ñ 
ô 
ˆŒˆˆr,   Nrˆ   Úposition_embeddingsr¥   Úpast_key_valuesrY   rL   c                 ó&  — |j         d d…         }g |¢d‘| j        ‘R }|                      |¦  «                             |¦  «                             dd¦  «        }|                      |¦  «                             |¦  «                             dd¦  «        }	|                      |¦  «                             |¦  «                             dd¦  «        }
|\  }}t          ||	||d¬¦  «        \  }}	|�|                     |	|
| j	        ¦  «        \  }	}
t          j        | j        j        t          ¦  «        } || ||	|
|f| j        sdn| j        | j        dœ|¤Ž\  }} |j        g |¢d‘R Ž                      ¦   «         }|                      |¦  «        }||fS )Nr^   r   r   )r•   r    )r§   r¦   )rg   rO   r¿   Úviewrk   rÀ   rÁ   r˜   Úupdaterº   r   Úget_interfacer6   Ú_attn_implementationr·   r¬   r¼   r¦   r›   r²   rÂ   )rI   rˆ   rÄ   r¥   rÅ   rY   Úinput_shapeÚhidden_shapeÚquery_statesr³   r´   rm   rn   Úattention_interfacer¶   rµ   s                   r-   ru   zXcodec2Attention.forward  sÈ  € ð $Ô)¨#¨2¨#Ô.ˆØ8˜Ð8 bÐ8¨$¬-Ð8Ð8ˆà—{’{ =Ñ1Ô1×6Ò6°|ÑDÔD×NÒNÈqÐRSÑTÔTˆØ—[’[ Ñ/Ô/×4Ò4°\ÑBÔB×LÒLÈQÐPQÑRÔRˆ
Ø—{’{ =Ñ1Ô1×6Ò6°|ÑDÔD×NÒNÈqÐRSÑTÔTˆà&‰ˆˆSõ $8¸ÀjÐRUÐWZÐjkÐ#lÑ#lÔ#lÑ ˆ�jàÐ&Ø'6×'=Ò'=¸jÈ,ÐX\ÔXfÑ'gÔ'gÑ$ˆJ˜å(?Ô(MØŒKÔ,Õ.Eñ)
ô )
Ðð %8Ð$7ØØØØØð	%
ð  $œ}ÐH�C�C°$Ô2HØ”Lð	%
ð 	%
ð ð	%
ð 	%
Ñ!ˆ�\ð *�kÔ)Ð;¨;Ð;¸Ð;Ð;Ð;×FÒFÑHÔHˆØ—k’k +Ñ.Ô.ˆØ˜LÐ(Ð(r,   )NNN)r"   r#   r$   r%   r   Úintr=   r&   r*   rx   r	   r   r   ru   rz   r{   s   @r-   r¹   r¹   ñ   så   ø€ € € € € àGÐGð
˜}ð 
¸ð 
ð 
ð 
ð 
ð 
ð 
ð4 IMØ.2Ø(,ð))ð ))à”|ð))ð # 5¤<°´Ð#=Ô>ÀÑEð))ð œ tÑ+ð	))ð
  ™ð))ð Ð+Ô,ð))ð 
ˆuŒ|˜Uœ\Ð)Ô	*ð))ð ))ð ))ð ))ð ))ð ))ð ))ð ))r,   r¹   ÚRMSNormc                   óT   ‡ — e Zd Zd	deddfˆ fd„Zdej        dej        fd„Zd„ Zˆ xZ	S )
ÚXcodec2RMSNormç�íµ ÷Æ°>ÚepsrL   Nc                 ó¬   •— t          ¦   «                              ¦   «          t          j        t	          j        |¦  «        ¦  «        | _        || _        dS )z=
        Xcodec2RMSNorm is equivalent to T5LayerNorm
        N)r<   r=   rD   r   r&   ÚonesÚweightÚvariance_epsilon)rI   rT   rÔ   rK   s      €r-   r=   zXcodec2RMSNorm.__init__:  sD   ø€ õ 	‰Œ×ÒÑÔÐÝ”l¥5¤:¨kÑ#:Ô#:Ñ;Ô;ˆŒØ #ˆÔÐÐr,   rˆ   c                 ó  — |j         }|                     t          j        ¦  «        }|                     d¦  «                             dd¬¦  «        }|t          j        || j        z   ¦  «        z  }| j        |                     |¦  «        z  S )Nr   r^   T)Úkeepdim)	rR   rX   r&   r±   ÚpowÚmeanÚrsqrtrØ   r×   )rI   rˆ   Úinput_dtypeÚvariances       r-   ru   zXcodec2RMSNorm.forwardB  s|   € Ø#Ô)ˆØ%×(Ò(­¬Ñ7Ô7ˆØ ×$Ò$ QÑ'Ô'×,Ò,¨R¸Ð,Ñ>Ô>ˆØ%­¬°H¸tÔ?TÑ4TÑ(UÔ(UÑUˆØŒ{˜]×-Ò-¨kÑ:Ô:Ñ:Ð:r,   c                 óH   — t          | j        j        ¦  «        › d| j        › �S )Nz, eps=)rx   r×   rg   rØ   )rI   s    r-   Ú
extra_reprzXcodec2RMSNorm.extra_reprI  s&   € Ý˜œÔ)Ñ*Ô*ÐIÐI°$Ô2GÐIÐIÐIr,   )rÓ   )
r"   r#   r$   rW   r=   r&   r*   ru   rá   rz   r{   s   @r-   rÒ   rÒ   8  sŒ   ø€ € € € € ð$ð $¨ð $¸$ð $ð $ð $ð $ð $ð $ð; U¤\ð ;°e´lð ;ð ;ð ;ð ;ðJð Jð Jð Jð Jð Jð Jr,   rÒ   c                   óÒ   ‡ — e Zd Zdedefˆ fd„Z	 	 	 	 	 ddej        dej        dz  dej        dz  d	e	dz  d
e
dz  deej        ej        f         dz  dee         dej        fd„Zˆ xZS )ÚXcodec2DecoderLayerr6   rº   c                 ó4  •— t          ¦   «                              ¦   «          |j        | _        t          ||¬¦  «        | _        t          |¦  «        | _        t          |j        |j        ¬¦  «        | _	        t          |j        |j        ¬¦  «        | _
        d S )N)r6   rº   ©rÔ   )r<   r=   rT   r¹   Ú	self_attnr}   ÚmlprÒ   Úrms_norm_epsÚinput_layernormÚpost_attention_layernormrÃ   s      €r-   r=   zXcodec2DecoderLayer.__init__N  sƒ   ø€ Ý‰Œ×ÒÑÔÐØ!Ô-ˆÔå)°À9ÐMÑMÔMˆŒå˜fÑ%Ô%ˆŒÝ-¨fÔ.@ÀfÔFYÐZÑZÔZˆÔÝ(6°vÔ7IÈvÔObÐ(cÑ(cÔ(cˆÔ%Ð%Ð%r,   NFrˆ   r¥   rp   rÅ   Ú	use_cacherÄ   rY   rL   c           
      óÎ   — |}|                       |¦  «        } | j        d||||||dœ|¤Ž\  }}	||z   }|}|                      |¦  «        }|                      |¦  «        }||z   }|S )N)rˆ   r¥   rp   rÅ   rë   rÄ   r+   )ré   ræ   rê   rç   )
rI   rˆ   r¥   rp   rÅ   rë   rÄ   rY   ÚresidualÚ_s
             r-   ru   zXcodec2DecoderLayer.forwardX  s¡   € ð !ˆØ×,Ò,¨]Ñ;Ô;ˆà)˜4œ>ð 
Ø'Ø)Ø%Ø+ØØ 3ð
ð 
ð ð
ð 
Ñˆ�qð ! =Ñ0ˆð !ˆØ×5Ò5°mÑDÔDˆØŸš Ñ/Ô/ˆØ  =Ñ0ˆØÐr,   )NNNFN)r"   r#   r$   r   rÏ   r=   r&   r*   r)   r	   Úboolrx   r   r   ru   rz   r{   s   @r-   rã   rã   M  sÿ   ø€ € € € € ðd˜}ð d¸ð dð dð dð dð dð dð /3Ø04Ø(,Ø!&ØHLðð à”|ðð œ tÑ+ðð Ô&¨Ñ-ð	ð
  ™ðð ˜$‘;ðð # 5¤<°´Ð#=Ô>ÀÑEðð Ð+Ô,ðð 
Œðð ð ð ð ð ð ð r,   rã   c                   ó*   ‡ — e Zd ZdZdˆ fd„	Zd„ Zˆ xZS )ÚXcodec2SnakeBetaa  
    A modified Snake function which uses separate parameters for the magnitude of the periodic components
    Shape:
        - Input: (B, C, T)
        - Output: (B, C, T), same shape as the input
    Parameters:
        - alpha - trainable parameter that controls frequency
        - beta - trainable parameter that controls magnitude
    References:
        - This activation function is a modified version based on this paper by Liu Ziyin, Tilman Hartwig, Masahito Ueda:
        https://huggingface.co/papers/2006.08195
    rP   c                 ó  •— t          ¦   «                              ¦   «          || _        t          t	          j        |¦  «        |z  ¦  «        | _        t          t	          j        |¦  «        |z  ¦  «        | _        d| _        d S )Ng•Ö&è.>)	r<   r=   Úin_featuresr   r&   ÚzerosÚalphaÚbetaÚno_div_by_zero)rI   ró   rõ   rK   s      €r-   r=   zXcodec2SnakeBeta.__init__†  sm   ø€ Ý‰Œ×ÒÑÔÐØ&ˆÔõ �uœ{¨;Ñ7Ô7¸%Ñ?Ñ@Ô@ˆŒ
Ý�eœk¨+Ñ6Ô6¸Ñ>Ñ?Ô?ˆŒ	à)ˆÔÐÐr,   c                 ó€  — | j                              d¦  «                             d¦  «        }| j                             d¦  «                             d¦  «        }t          j        |¦  «        }t          j        |¦  «        }|d|| j        z   z  t          j        t          j        ||z  ¦  «        d¦  «        z  z   }|S )u’   
        Forward pass of the function.
        Applies the function to the input elementwise.
        SnakeBeta âˆ¶= x + 1/b * sin^2 (xa)
        r   r^   rP   r   )rõ   r’   rö   r&   Úexpr÷   rÛ   rn   )rI   rˆ   rõ   rö   s       r-   ru   zXcodec2SnakeBeta.forward�  s°   € ð ”
×$Ò$ QÑ'Ô'×1Ò1°"Ñ5Ô5ˆØŒy×"Ò" 1Ñ%Ô%×/Ò/°Ñ3Ô3ˆÝ”	˜%Ñ Ô ˆÝŒy˜‰ŒˆØ%¨°°tÔ7JÑ0JÑ)KÍuÌyÝŒI�m eÑ+Ñ,Ô,¨añP
ô P
ñ )
ñ 
ˆð Ðr,   )rP   ©r"   r#   r$   r%   r=   ru   rz   r{   s   @r-   rñ   rñ   x  sV   ø€ € € € € ðð ð*ð *ð *ð *ð *ð *ðð ð ð ð ð ð r,   rñ   c                 óX  — |dz  dk    }|dz  }d|z  }d|dz
  z  t           j        z  |z  dz   }|dk    r	d|d	z
  z  }n|d
k    rd|dz
  dz  z  d|d
z
  z  z   }nd}t          j        ||dt          j        ¬¦  «        }|rt          j        | |¦  «        dz   }	nt          j        |¦  «        |z
  }	| dk    r#t          j        dd|ft          j        ¬¦  «        S t          j        d| z  |	z  ¦  «        }
d| z  |z  |
z  }||                     ¦   «         z  }| 	                    dd|¦  «        S )aB  Generates a 1D Kaiser-windowed sinc filter.

    Args:
        cutoff (float): Normalized cutoff frequency (0 to 0.5).
        half_width (float): Transition bandwidth.
        kernel_size (int): Number of filter taps.

    Returns:
        torch.Tensor: A tensor of shape (1, 1, kernel_size) representing the filter.
    r   r   é   gHáz®G@r   gÍÌÌÌÌÌ@g      I@gKê46¼?gffffff!@g      5@g¨WÊ2Ä±â?é   gš™™™™™Ù?gUjö@+0´?r    F)rö   ÚperiodicrR   ç      à?rQ   )
ÚmathÚpir&   Úkaiser_windowr±   rV   rô   ÚsincÚsumrÇ   )ÚcutoffÚ
half_widthÚkernel_sizeÚis_evenÚ	half_sizeÚdelta_fÚattenuationrö   r  Útime_indicesÚsinc_filterÚnormalized_filters               r-   Úkaiser_sinc_filter1dr  ¡  ss  € ð ˜A‰o Ò"€GØ˜qÑ €Ið �*‰n€GØ˜9 q™=Ñ)­D¬GÑ3°gÑ=ÀÑD€Kà�TÒÐØ˜ sÑ*Ñ+ˆˆØ	˜Ò	Ð	Ø˜ rÑ)¨cÑ1Ñ1°G¸{ÈTÑ?QÑ4RÑRˆˆàˆåÔ'¨¸$ÈÕV[ÔVcÐdÑdÔd€Mð ð =Ý”| Y J°	Ñ:Ô:¸SÑ@ˆˆå”| KÑ0Ô0°9Ñ<ˆð �‚{€{ÝŒ{˜A˜q +Ð.µe´mÐDÑDÔDÐDå”*˜Q ™Z¨,Ñ6Ñ7Ô7€KØ˜F™
 ]Ñ2°[Ñ@Ðð Ð*×.Ò.Ñ0Ô0Ñ0Ðà×!Ò! ! Q¨Ñ4Ô4Ð4r,   c                   ó&   ‡ — e Zd Zdˆ fd„	Zd„ Zˆ xZS )ÚXcodec2DownSample1dr   Nc                 ó®  •— t          ¦   «                              ¦   «          d|z  }d|z  }|| _        || _        || _        |dk     rt          d¦  «        ‚|dk    rt          d¦  «        ‚|dz  dk    | _        |dz  t          | j        ¦  «        z
  | _        |dz  | _	        || _
        t          |||¦  «        }t          j        |d¬	¦  «        | _        d S )
Nrÿ   ç333333ã?r    z(Minimum cutoff must be larger than zero.z'A cutoff above 0.5 does not make sense.r   r   Fr:   )r<   r=   r  r  r  Ú
ValueErrorÚevenrÏ   Úpad_leftÚ	pad_rightÚstrider  rD   rE   Úfilter)rI   Úratior  r  r  r  rK   s         €r-   r=   zXcodec2DownSample1d.__init__Ð  sØ   ø€ Ý‰Œ×ÒÑÔÐØ�u‘ˆØ˜5‘[ˆ
ØˆŒØ$ˆŒØ&ˆÔà�CŠ<ˆ<ÝÐGÑHÔHÐHØ�CŠ<ˆ<ÝÐFÑGÔGÐGà !‘O qÒ(ˆŒ	Ø# qÑ(­3¨t¬y©>¬>Ñ9ˆŒØ$¨Ñ)ˆŒØˆŒÝ% f¨j¸+ÑFÔFˆÝ”i °5Ð9Ñ9Ô9ˆŒˆˆr,   c                 ó  — |j         d         }t          j        || j        | j        fd¬¦  «        }t          j        || j                             |j        ¦  «         	                    |dd¦  «        | j
        |¬¦  «        }|S )Nr   Ú	replicate©Úmoder^   ©r  Úgroups)rg   ÚFÚpadr  r  Úconv1dr  rX   rR   rf   r  )rI   rˆ   ÚchannelsÚouts       r-   ru   zXcodec2DownSample1d.forwardä  s}   € Ø Ô& qÔ)ˆÝœ˜m¨d¬m¸T¼^Ð-LÐS^Ð_Ñ_Ô_ˆÝŒhØàŒK�NŠN˜=Ô.Ñ/Ô/×6Ò6°xÀÀRÑHÔHØ”;Øð
ñ 
ô 
ˆð ˆ
r,   ©r   N©r"   r#   r$   r=   ru   rz   r{   s   @r-   r  r  Ï  sL   ø€ € € € € ð:ð :ð :ð :ð :ð :ð(
ð 
ð 
ð 
ð 
ð 
ð 
r,   r  c                   ó&   ‡ — e Zd Zdˆ fd„	Zd„ Zˆ xZS )ÚXcodec2UpSample1dr   Nc                 óÜ  •— t          ¦   «                              ¦   «          || _        |€t          d|z  dz  ¦  «        dz  n|| _        || _        | j        |z  dz
  | _        | j        | j        z  | j        | j        z
  dz  z   | _        | j        | j        z  | j        | j        z
  dz   dz  z   | _        t          d|z  d|z  | j        ¬¦  «        }t          j        |d¬¦  «        | _        d S )	Né   r   r   rÿ   r  )r  r  r  Fr:   )r<   r=   r  rÏ   r  r  r"  r  r  r  rD   rE   r  )rI   r  r  r  rK   s       €r-   r=   zXcodec2UpSample1d.__init__ò  sì   ø€ Ý‰Œ×ÒÑÔÐØˆŒ
Ø6AÐ6I�3˜q 5™y¨A™~Ñ.Ô.°Ñ2Ð2È{ˆÔØˆŒØÔ# uÑ,¨qÑ0ˆŒØœ 4¤;Ñ.°$Ô2BÀTÄ[Ñ2PÐUVÑ1VÑVˆŒØœ D¤KÑ/°4Ô3CÀdÄkÑ3QÐTUÑ3UÐZ[Ñ2[Ñ[ˆŒå%¨S°5©[ÀSÈ5Á[Ð^bÔ^nÐoÑoÔoˆÝ”i °5Ð9Ñ9Ô9ˆŒˆˆr,   c           	      óB  — |j         d         }t          j        || j        | j        fd¬¦  «        }| j        t          j        || j                             |j        ¦  «                             |dd¦  «        | j	        |¬¦  «        z  }|d| j
        | j         …f         }|S )Nr   r  r  r^   r  .)rg   r!  r"  r  Úconv_transpose1dr  rX   rR   rf   r  r  r  )rI   rˆ   r$  s      r-   ru   zXcodec2UpSample1d.forwardþ  s¡   € Ø Ô& qÔ)ˆÝœ˜m¨d¬h¸¼Ð-AÈÐTÑTÔTˆØœ
¥QÔ%7ØàŒK�NŠN˜=Ô.Ñ/Ô/×6Ò6°xÀÀRÑHÔHØ”;Øð&
ñ &
ô &
ñ 
ˆð & c¨4¬=¸D¼N¸?Ð+JÐ&JÔKˆØÐr,   r&  r'  r{   s   @r-   r)  r)  ñ  sL   ø€ € € € € ð
:ð 
:ð 
:ð 
:ð 
:ð 
:ðð ð ð ð ð ð r,   r)  c            	       ó@   ‡ — e Zd Z	 	 	 	 d	dedededefˆ fd„Zd„ Zˆ xZS )
ÚXcodec2AntiAliasedActivation1dr   é   Úup_ratioÚ
down_ratioÚup_kernel_sizeÚdown_kernel_sizec                 óæ   •— t          ¦   «                              ¦   «          t          |¦  «        st          d¦  «        ‚|| _        t          ||¦  «        | _        t          ||¦  «        | _        d S )Nz$Activation function must be callable)	r<   r=   ÚcallableÚ	TypeErrorÚactr)  Úupsampler  Ú
downsample)rI   Ú
activationr1  r2  r3  r4  rK   s         €r-   r=   z'Xcodec2AntiAliasedActivation1d.__init__  si   ø€ õ 	‰Œ×ÒÑÔÐÝ˜
Ñ#Ô#ð 	DÝÐBÑCÔCÐCØˆŒÝ)¨(°NÑCÔCˆŒÝ-¨jÐ:JÑKÔKˆŒˆˆr,   c                 ó„   — |                       |¦  «        }|                      |¦  «        }|                      |¦  «        }|S rv   )r9  r8  r:  rŠ   s     r-   ru   z&Xcodec2AntiAliasedActivation1d.forward  s;   € ØŸš mÑ4Ô4ˆØŸš Ñ/Ô/ˆØŸš¨Ñ6Ô6ˆàÐr,   )r   r   r0  r0  )r"   r#   r$   rÏ   r=   ru   rz   r{   s   @r-   r/  r/    s’   ø€ € € € € ð ØØ Ø "ðLð Lð ðLð ð	Lð
 ðLð ðLð Lð Lð Lð Lð Lðð ð ð ð ð ð r,   r/  c                   ó(   ‡ — e Zd ZdZˆ fd„Zd„ Zˆ xZS )ÚXcodec2ResidualUnitza
    A residual unit composed of Snake1d and weight-normalized Conv1d layers with dilations.
    c                 óT  •— t          ¦   «                              ¦   «          d|z  dz  }t          t          |¦  «        ¬¦  «        | _        t          j        ||d||¬¦  «        | _        t          t          |¦  «        ¬¦  «        | _        t          j        ||d¬¦  «        | _	        d S )Nr+  r   ©r;  é   )r  ÚdilationÚpaddingr   )r  )
r<   r=   r/  rñ   Úsnake1rD   ÚConv1dÚconv1Úsnake2Úconv2)rI   Ú	dimensionrB  r"  rK   s       €r-   r=   zXcodec2ResidualUnit.__init__)  s™   ø€ Ý‰Œ×ÒÑÔÐØ˜Ñ! aÑ'ˆÝ4Õ@PÐQZÑ@[Ô@[Ð\Ñ\Ô\ˆŒÝ”Y˜y¨)ÀÈXÐ_bÐcÑcÔcˆŒ
Ý4Õ@PÐQZÑ@[Ô@[Ð\Ñ\Ô\ˆŒÝ”Y˜y¨)ÀÐCÑCÔCˆŒ
ˆ
ˆ
r,   c                 ó  — |}|                       |                      |¦  «        ¦  «        }|                      |                      |¦  «        ¦  «        }|j        d         |j        d         z
  dz  }|dk    r|d|| …f         }||z   }|S )ar  
        Forward pass through the residual unit.

        Args:
            hidden_state (`torch.Tensor` of shape `(batch_size, channels, time_steps)`):
                Input tensor .

        Returns:
            output_tensor (`torch.Tensor` of shape `(batch_size, channels, time_steps)`):
                Input tensor after passing through the residual unit.
        r^   r   r   .)rF  rD  rH  rG  rg   )rI   Úhidden_stateÚoutput_tensorrC  s       r-   ru   zXcodec2ResidualUnit.forward1  s‘   € ð %ˆØŸ
š
 4§;¢;¨}Ñ#=Ô#=Ñ>Ô>ˆØŸ
š
 4§;¢;¨}Ñ#=Ô#=Ñ>Ô>ˆàÔ% bÔ)¨MÔ,?ÀÔ,CÑCÈÑIˆØ�QŠ;ˆ;Ø'¨¨W°g°XÐ-=Ð(=Ô>ˆLØ$ }Ñ4ˆØÐr,   rú   r{   s   @r-   r>  r>  $  sV   ø€ € € € € ðð ðDð Dð Dð Dð Dðð ð ð ð ð ð r,   r>  c                   ó8   ‡ — e Zd ZdZddededefˆ fd„Zd„ Zˆ xZS )	ÚXcodec2EncoderBlockz&Encoder block used in XCODEC2 encoder.r   r6   r  Ústride_indexc           
      ó´  •— t          ¦   «                              ¦   «          |j        d|z  z  }t          |dz  d¬¦  «        | _        t          |dz  d¬¦  «        | _        t          |dz  d¬¦  «        | _        t          t          |dz  ¦  «        ¬¦  «        | _	        t          j        |dz  |d|z  |t          j        |dz  ¦  «        ¬¦  «        | _        d S )Nr   r   )rB  r   é	   r@  ©r  r  rC  )r<   r=   Úencoder_hidden_sizer>  Ú	res_unit1Ú	res_unit2Ú	res_unit3r/  rñ   rD  rD   rE  r   ÚceilrF  )rI   r6   r  rO  rI  rK   s        €r-   r=   zXcodec2EncoderBlock.__init__K  sÖ   ø€ Ý‰Œ×ÒÑÔÐØÔ.°°L±Ñ@ˆ	Ý,¨Y¸!©^ÀaÐHÑHÔHˆŒÝ,¨Y¸!©^ÀaÐHÑHÔHˆŒÝ,¨Y¸!©^ÀaÐHÑHÔHˆŒÝ4Õ@PÐQZÐ^_ÑQ_Ñ@`Ô@`ÐaÑaÔaˆŒÝ”YØ˜‰N˜I°1°v±:ÀfÕVZÔV_Ð`fÐijÑ`jÑVkÔVkð
ñ 
ô 
ˆŒ
ˆ
ˆ
r,   c                 óÔ   — |                       |¦  «        }|                      |¦  «        }|                      |                      |¦  «        ¦  «        }|                      |¦  «        }|S rv   )rT  rU  rD  rV  rF  )rI   rK  s     r-   ru   zXcodec2EncoderBlock.forwardV  sX   € Ø—~’~ lÑ3Ô3ˆØ—~’~ lÑ3Ô3ˆØ—{’{ 4§>¢>°,Ñ#?Ô#?Ñ@Ô@ˆØ—z’z ,Ñ/Ô/ˆàÐr,   )r   r   )	r"   r#   r$   r%   r   rÏ   r=   ru   rz   r{   s   @r-   rN  rN  H  sl   ø€ € € € € Ø0Ð0ð	
ð 	
˜}ð 	
°cð 	
ÈSð 	
ð 	
ð 	
ð 	
ð 	
ð 	
ðð ð ð ð ð ð r,   rN  c                   ó.   ‡ — e Zd ZdZdefˆ fd„Zd„ Zˆ xZS )ÚXcodec2EncoderzXCODEC2 Encoderr6   c                 ó  •— t          ¦   «                              ¦   «          t          j        d|j        dd¬¦  «        | _        g | _        t          |j        ¦  «        D ]+\  }}|dz   }| xj        t          |||¬¦  «        gz  c_        Œ,t          j
        | j        ¦  «        | _        |j        dt          |j        ¦  «        z  z  }t          t          |¦  «        ¬¦  «        | _        t          j        ||j        dd¬¦  «        | _        d S )Nr   rA  r   ©r  rC  )r  rO  r   r@  )r<   r=   rD   rE  rS  rF  ÚblockÚ	enumerateÚdownsampling_ratiosrN  Ú
ModuleListÚlenr/  rñ   rD  rT   rH  )rI   r6   rO  r  Úd_modelrK   s        €r-   r=   zXcodec2Encoder.__init__b  sÿ   ø€ Ý‰Œ×ÒÑÔÐõ ”Y˜q &Ô"<È!ÐUVÐWÑWÔWˆŒ
àˆŒ
å$-¨fÔ.HÑ$IÔ$Ið 	bð 	bÑ ˆL˜&Ø'¨!Ñ+ˆLØˆJŒJÕ.¨v¸fÐS_Ð`Ñ`Ô`ÐaÑaˆJŒJˆJå”] 4¤:Ñ.Ô.ˆŒ
ØÔ,¨qµC¸Ô8RÑ4SÔ4SÑ/SÑSˆÝ4Õ@PÐQXÑ@YÔ@YÐZÑZÔZˆŒÝ”Y˜w¨Ô(:ÈÐSTÐUÑUÔUˆŒ
ˆ
ˆ
r,   c                 ó®   — |                       |¦  «        }| j        D ]} ||¦  «        }Œ|                      |¦  «        }|                      |¦  «        }|S rv   )rF  r]  rD  rH  )rI   rK  r¡   s      r-   ru   zXcodec2Encoder.forwards  s]   € Ø—z’z ,Ñ/Ô/ˆà”jð 	0ð 	0ˆFØ!˜6 ,Ñ/Ô/ˆLˆLà—{’{ <Ñ0Ô0ˆØ—z’z ,Ñ/Ô/ˆàÐr,   )r"   r#   r$   r%   r   r=   ru   rz   r{   s   @r-   rZ  rZ  _  s`   ø€ € € € € ØÐðV˜}ð Vð Vð Vð Vð Vð Vð"	ð 	ð 	ð 	ð 	ð 	ð 	r,   rZ  c                   óH   ‡ — e Zd Zdefˆ fd„Zdej        dej        fd„Zˆ xZS )ÚXcodec2ResNetBlockr6   c                 óè  •— t          ¦   «                              ¦   «          t          j        d|j        dd¬¦  «        | _        t          j        ¦   «         | _        t          j        |j        |j        ddd¬¦  «        | _	        t          j        d|j        dd¬¦  «        | _
        t          j        ¦   «         | _        |j        | _        t          j        |j        |j        ddd¬¦  «        | _        d S )Né    rÓ   T)Ú
num_groupsÚnum_channelsrÔ   Úaffiner   r   rR  )r<   r=   rD   Ú	GroupNormrT   Únorm1ÚSiLUÚactivation1rE  rF  Únorm2Úactivation2Úactivation_dropoutrH  r‡   s     €r-   r=   zXcodec2ResNetBlock.__init__€  sÌ   ø€ Ý‰Œ×ÒÑÔÐÝ”\¨R¸fÔ>PÐVZÐcgÐhÑhÔhˆŒ
Ýœ7™9œ9ˆÔÝ”Y˜vÔ1°6Ô3EÐSTÐ]^ÐhiÐjÑjÔjˆŒ
Ý”\¨R¸fÔ>PÐVZÐcgÐhÑhÔhˆŒ
Ýœ7™9œ9ˆÔØ"(Ô";ˆÔÝ”Y˜vÔ1°6Ô3EÐSTÐ]^ÐhiÐjÑjÔjˆŒ
ˆ
ˆ
r,   rˆ   rL   c                 ó¸  — |                      dd¦  «        }|}|                      |¦  «        }|                      |¦  «        }|                      |¦  «        }|                      |¦  «        }|                      |¦  «        }t          j                             || j	        | j
        ¬¦  «        }|                      |¦  «        }||z                         dd¦  «        S )Nr   r   rª   )rk   rl  rn  rF  ro  rp  rD   r¯   r§   rq  r¬   rH  ©rI   rˆ   rí   s      r-   ru   zXcodec2ResNetBlock.forwardŠ  sÄ   € Ø%×/Ò/°°1Ñ5Ô5ˆØ ˆØŸ
š
 =Ñ1Ô1ˆØ×(Ò(¨Ñ7Ô7ˆØŸ
š
 =Ñ1Ô1ˆØŸ
š
 =Ñ1Ô1ˆØ×(Ò(¨Ñ7Ô7ˆÝœ×-Ò-¨m¸tÔ?VÐaeÔanÐ-ÑoÔoˆØŸ
š
 =Ñ1Ô1ˆØ Ñ(×3Ò3°A°qÑ9Ô9Ð9r,   r‹   r{   s   @r-   re  re    sq   ø€ € € € € ðk˜}ð kð kð kð kð kð kð
: U¤\ð 
:°e´lð 
:ð 
:ð 
:ð 
:ð 
:ð 
:ð 
:ð 
:r,   re  c                   ó¼   ‡ — e Zd ZdZdefˆ fd„Zdd„Zdej        dej        fd„Z	dd
ej        de
dej        fd„Zd
ej        deej        ej        f         fd„Zˆ xZS )ÚXcodec2FiniteScalarQuantizationa!  
    Finite Scalar Quantization (FSQ) module that quantizes continuous latent representations into discrete codes.
    Original code: https://github.com/lucidrains/vector-quantize-pytorch/blob/353d46027888dfb140c3c65a67a7356f1492d71d/vector_quantize_pytorch/finite_scalar_quantization.py#L64

    Original modeling uses `ResidualFSQ` with a single quantizer: https://huggingface.co/HKUSTAudio/xcodec2/blob/main/vq/codec_decoder_vocos.py#L389
    But we can directly use FSQ since a main feature of Xcodec2 is that it uses a single codebook.
    r6   c                 óL  •— t          ¦   «                              ¦   «          t          |j        ¦  «        | _        |                      ¦   «         \  }}}t          j        |d¬¦  «        | _        t          j        |d¬¦  «        | _        t          j        |d¬¦  «        | _	        d S )NFr:   )
r<   r=   ÚlistÚquantization_levelsÚ_compute_buffersrD   rE   ÚlevelsÚbasisÚcodebook)rI   r6   rz  r{  r|  rK   s        €r-   r=   z(Xcodec2FiniteScalarQuantization.__init__   sŠ   ø€ Ý‰Œ×ÒÑÔÐÝ#'¨Ô(BÑ#CÔ#CˆÔ Ø"&×"7Ò"7Ñ"9Ô"9Ñˆ��xÝ”i °5Ð9Ñ9Ô9ˆŒÝ”Y˜u°Ð7Ñ7Ô7ˆŒ
Ýœ	 (°uÐ=Ñ=Ô=ˆŒˆˆr,   Nc                 ó¨  — t          j        | j        t           j        |¬¦  «        }t          j        t          j        dg| j        dd…         z   |¬¦  «        dt           j        ¬¦  «        }t          j        t          t          j        | j        ¦  «        ¦  «        |¬¦  «         	                    d¦  «        }||z  |z  }|dz  }||z
  |z  }|||fS )	zFCompute the levels, basis, and codebook buffers for the FSQ quantizer.r_   r   Nr^   ©r4   r   r©   r   )
r&   Útensorrx  Úint32ÚcumprodrV   rÏ   ÚnpÚprodr’   )rI   r4   rz  r{  ÚindicesÚlevel_indicesr  r|  s           r-   ry  z0Xcodec2FiniteScalarQuantization._compute_buffers¨  sÐ   € å”˜dÔ6½e¼kÐRXÐYÑYÔYˆÝ”ÝŒL˜!˜˜tÔ7¸¸¸Ô<Ñ<ÀVÐLÑLÔLÐRSÕ[`Ô[fð
ñ 
ô 
ˆõ ”,�s¥2¤7¨4Ô+CÑ#DÔ#DÑEÔEÈfÐUÑUÔU×_Ò_Ð`bÑcÔcˆØ  EÑ)¨VÑ3ˆØ˜q‘[ˆ
Ø! JÑ.°*Ñ<ˆØ�u˜hÐ&Ð&r,   r„  rL   c                 óx   — |                      d¦  «        }|| j        z  | j        z  }| j        dz  }||z
  |z  }|S )z`
        Convert integer codebook indices to normalized per-dimension codes in [-1, 1].
        r^   r   )r’   r{  rz  )rI   r„  r…  r  Úcodess        r-   Ú_indices_to_codesz1Xcodec2FiniteScalarQuantization._indices_to_codes´  sJ   € ð ×#Ò# BÑ'Ô'ˆØ  D¤JÑ.°$´+Ñ=ˆØ”[ AÑ%ˆ
Ø Ñ+¨zÑ9ˆØˆr,   çü©ñÒMbP?rˆ   rÔ   c                 óÔ   — | j         dz
  d|z   z  dz  }t          j        | j         dz  dk    dd¦  «        }||z                       ¦   «         }||z                        ¦   «         |z  |z
  S )að  
        Constrain `hidden_states` to the valid quantization range for each dimension.

        Uses a scaled tanh to soft-clip values into the interval
        $[-(L-1)/2, (L-1)/2]$ (offset by 0.5 for even-level dimensions), where $L$ is
        the number of quantization levels. The small `eps` margin prevents values from
        saturating exactly at the boundary, which would zero out gradients.

        Args:
            hidden_states (`torch.Tensor`): Continuous input to be bounded.
            eps (`float`, *optional*, defaults to `1e-3`):
                Small margin added to the level range to avoid gradient saturation at boundaries.

        Returns:
            `torch.Tensor`: Bounded values in the valid quantization range.
        r   r   r   rÿ   r    )rz  r&   ÚwhereÚatanhÚtanh)rI   rˆ   rÔ   Ú
half_rangeÚoffsetÚshifts         r-   Úboundz%Xcodec2FiniteScalarQuantization.bound¾  sr   € ð" ”k A‘o¨!¨c©'Ñ2°QÑ6ˆ
Ý”˜Tœ[¨1™_°Ò1°3¸Ñ<Ô<ˆØ˜*Ñ$×+Ò+Ñ-Ô-ˆØ Ñ%×+Ò+Ñ-Ô-°
Ñ:¸VÑCÐCr,   c                 ó\  — |j         }t          |j        j        t          ¦  «        r|j        j        dk    r|j        j        nd}t          |d¬¦  «        5  |                     ¦   «         }| j        dz  }|                      |¦  «        }| 	                    ¦   «         }|||z
   
                    ¦   «         z   }||z  }||z  |z   }|| j        z                       d¬¦  «                             t          j        ¦  «        }d d d ¦  «         n# 1 swxY w Y   |                     |¦  «        |fS )Nr`   ra   Frb   r   r^   re   )rR   rh   r4   ri   rj   r   rW   rz  r‘  ÚroundÚdetachr{  r  rX   r&   r€  )	rI   rˆ   Úoriginal_dtyperc   r  Úroundedr‡  Úcode_scaledr„  s	            r-   ru   z'Xcodec2FiniteScalarQuantization.forwardÔ  s€  € à&Ô,ˆõ ˜-Ô.Ô3µSÑ9Ô9ðØ>KÔ>RÔ>WÐ[`Ò>`Ð>`ð Ô Ô%Ð%àð 	õ
 ¨¸UÐCÑCÔCð 
	Mð 
	MØ)×/Ò/Ñ1Ô1ˆMØœ¨Ñ)ˆJà ŸJšJ }Ñ5Ô5ˆMØ#×)Ò)Ñ+Ô+ˆGØ! W¨}Ñ%<×$DÒ$DÑ$FÔ$FÑFˆEØ˜JÑ&ˆEà  :Ñ-°Ñ;ˆKØ" T¤ZÑ/×4Ò4¸Ð4Ñ<Ô<×?Ò?ÅÄÑLÔLˆGð
	Mð 
	Mð 
	Mñ 
	Mô 
	Mð 
	Mð 
	Mð 
	Mð 
	Mð 
	Mð 
	Møøøð 
	Mð 
	Mð 
	Mð 
	Mð �xŠx˜Ñ'Ô'¨Ð0Ð0s   ÁB*DÄDÄDrv   )r‰  )r"   r#   r$   r%   r   r=   ry  r&   r*   rˆ  rW   r‘  rx   ru   rz   r{   s   @r-   ru  ru  —  sô   ø€ € € € € ðð ð>˜}ð >ð >ð >ð >ð >ð >ð
'ð 
'ð 
'ð 
'ð¨¬ð ¸%¼,ð ð ð ð ðDð D 5¤<ð D°eð DÀuÄ|ð Dð Dð Dð Dð,1 U¤\ð 1°e¸E¼LÈ%Ì,Ð<VÔ6Wð 1ð 1ð 1ð 1ð 1ð 1ð 1ð 1r,   ru  c                   óL   ‡ — e Zd ZdZdefˆ fd„Zdej        dej        fd„Zˆ xZ	S )ÚXcodec2ISTFTHeadzù
    Head for converting decoder outputs to waveform via STFT projection and ISTFT.

    Uses custom "same" padding ISTFT from Vocos:
    https://github.com/gemelo-ai/vocos/blob/c859e3b7b534f3776a357983029d34170ddd6fc3/vocos/spectral_ops.py#L47
    r6   c                 ó\  •— t          ¦   «                              ¦   «          t          j        |j        |j        dz   ¦  «        | _        |j        | _        |j        | _        | j        | j        z
  dz  | _        t          j
        |j        ¦  «        }t          j        |d¬¦  «        | _        d S )Nr   Fr:   )r<   r=   rD   rƒ   rT   Ún_fftÚlinearÚ
hop_lengthrC  r&   Úhann_windowrE   Úwindow)rI   r6   rŸ  rK   s      €r-   r=   zXcodec2ISTFTHead.__init__ò  sŠ   ø€ Ý‰Œ×ÒÑÔÐÝ”i Ô 2°F´LÀ1Ñ4DÑEÔEˆŒØ”\ˆŒ
Ø Ô+ˆŒØœ
 T¤_Ñ4¸Ñ:ˆŒÝÔ" 6¤<Ñ0Ô0ˆÝ”i °5Ð9Ñ9Ô9ˆŒˆˆr,   rˆ   rL   c                 ó  — |                       |¦  «        }|                     dd¬¦  «        \  }}|                     ¦   «         }|                     ¦   «         }t          j        |¦  «                             d¬¦  «        }t          j        ||¦  «        }t          j                             || j	        dd¬¦  «         
                    dd¦  «        }|| j        d d d …d f         z  }|j        d         }|dz
  | j        z  | j	        z   }t          j        |d|fd| j	        fd| j        f¬	¦  «        d d …d
d
| j        | j         …f         }	t          j        | j                             ¦   «                              d|d¦  «         
                    dd¦  «        d|fd| j	        fd| j        f¬	¦  «                             ¦   «         | j        | j         …         }
|
                     d¬¦  «        }
|	|
z  }	|	                     d¦  «        S )Nr   r^   re   g      Y@)ÚmaxÚbackward)r[   Únormr   )Úoutput_sizer  r  r   g•dyáý¥=)Úmin)rœ  ÚchunkrW   r&   rù   ÚclampÚpolarÚfftÚirfftr›  rk   rŸ  rg   r�  r!  ÚfoldrC  Úsquarerf   Úsqueezer’   )rI   rˆ   Ú	stft_predÚ	magnitudeÚphaseÚspectrogram_complexÚtime_framesÚ
num_framesr¤  ÚaudioÚwindow_envelopes              r-   ru   zXcodec2ISTFTHead.forwardû  sú  € Ø—K’K Ñ.Ô.ˆ	à$Ÿ?š?¨1°"˜?Ñ5Ô5Ñˆ	�5à—O’OÑ%Ô%ˆ	Ø—’‘”ˆå”I˜iÑ(Ô(×.Ò.°3Ð.Ñ7Ô7ˆ	Ý#œk¨)°UÑ;Ô;Ðõ ”i—o’oÐ&9¸4¼:È2ÐT^�oÑ_Ô_×iÒiÐjkÐmnÑoÔoˆØ! D¤K°°a°a°a¸°Ô$>Ñ>ˆØ(Ô.¨qÔ1ˆ
Ø! A‘~¨¬Ñ8¸4¼:ÑEˆÝ”ØØ˜KÐ(Ø˜DœJ˜Ø�t”Ð'ð	
ñ 
ô 
ð
 ˆ!ˆ!ˆQ��4”< 4¤< -Ð/Ð
/ô1ˆõ œ&ØŒK×ÒÑ Ô ×'Ò'¨¨:°rÑ:Ô:×DÒDÀQÈÑJÔJØ˜KÐ(Ø˜DœJ˜Ø�t”Ð'ð	
ñ 
ô 
÷
 Š'‰)Œ)�D”L D¤L =Ð0ô2ˆð *×/Ò/°EÐ/Ñ:Ô:ˆØ˜Ñ'ˆØ�Š˜qÑ!Ô!Ð!r,   ©
r"   r#   r$   r%   r   r=   r&   r*   ru   rz   r{   s   @r-   r™  r™  ê  st   ø€ € € € € ðð ð:˜}ð :ð :ð :ð :ð :ð :ð"" U¤\ð ""°e´lð ""ð ""ð ""ð ""ð ""ð ""ð ""ð ""r,   r™  c                   ó†   ‡ — e Zd Zdefˆ fd„Zdej        dej        fd„Zdej        deej        ej        f         fd„Z	ˆ xZ
S )ÚXcodec2Quantizerr6   c                 ó4  •— t          ¦   «                              ¦   «          t          |¦  «        | _        t	          j        |j        t          |j        ¦  «        ¦  «        | _	        t	          j        t          |j        ¦  «        |j        ¦  «        | _
        d S rv   )r<   r=   ru  Ú	quantizerrD   rƒ   Úquantization_dimra  rx  Ú
project_inÚproject_outr‡   s     €r-   r=   zXcodec2Quantizer.__init__!  sq   ø€ Ý‰Œ×ÒÑÔÐÝ8¸Ñ@Ô@ˆŒÝœ) FÔ$;½SÀÔA[Ñ=\Ô=\Ñ]Ô]ˆŒÝœ9¥S¨Ô)CÑ%DÔ%DÀfÔF]Ñ^Ô^ˆÔÐÐr,   r„  rL   c                 óz   — |                      d¦  «        }| j        j        |         }|                      |¦  «        S ©Nr^   )r­  rº  r|  r½  )rI   r„  r‡  s      r-   Ú
from_codeszXcodec2Quantizer.from_codes'  s6   € Ø—/’/ "Ñ%Ô%ˆØ”Ô'¨Ô0ˆØ×Ò Ñ&Ô&Ð&r,   rˆ   c                 ó   — |                       |¦  «        }|j        }| j                             |¦  «        }|                      |¦  «        \  }}|                      |                     |¦  «        ¦  «        }|                     d¦  «        }||fS r¿  )r¼  rR   rº  r‘  r½  rX   r’   )rI   rˆ   r•  Úquantized_outr„  s        r-   ru   zXcodec2Quantizer.forward,  s…   € ØŸš¨Ñ6Ô6ˆØ&Ô,ˆØœ×,Ò,¨]Ñ;Ô;ˆØ!%§¢°Ñ!>Ô!>Ñˆ�wØ×(Ò(¨×)9Ò)9¸.Ñ)IÔ)IÑJÔJˆØ×#Ò# BÑ'Ô'ˆØ˜gÐ%Ð%r,   )r"   r#   r$   r   r=   r&   r*   rÀ  rx   ru   rz   r{   s   @r-   r¸  r¸     s£   ø€ € € € € ð_˜}ð _ð _ð _ð _ð _ð _ð' %¤,ð '°5´<ð 'ð 'ð 'ð 'ð
& U¤\ð &°e¸E¼LÈ%Ì,Ð<VÔ6Wð &ð &ð &ð &ð &ð &ð &ð &r,   r¸  c                   óL   ‡ — e Zd ZdZdefˆ fd„Zdej        dej        fd„Zˆ xZ	S )ÚXcodec2DecoderzVVocos-based decoder with ResNet, Transformer, and ISTFT head for audio reconstruction.r6   c                 óæ  •‡— t          ¦   «                              ¦   «          t          j        ‰j        ‰j        j        z   ‰j        ¦  «        | _        t          j        ‰j        ‰j        dd¬¦  «        | _        t          j	        t          ‰¦  «        t          ‰¦  «        g¦  «        | _        ‰j        | _        t          ‰¬¦  «        | _        t          j	        ˆfd„t          ‰j        ¦  «        D ¦   «         ¦  «        | _        t          j	        t          ‰¦  «        t          ‰¦  «        g¦  «        | _        t          j        ‰j        d¬¦  «        | _        t+          ‰¦  «        | _        d S )NrA  r   r\  )r6   c                 ó0   •— g | ]}t          ‰|¦  «        ‘ŒS r+   )rã   )Ú.0rº   r6   s     €r-   ú
<listcomp>z+Xcodec2Decoder.__init__.<locals>.<listcomp>A  s$   ø€ ÐeÐeÐe¸	Õ  ¨Ñ3Ô3ÐeÐeÐer,   rÓ   rå   )r<   r=   rD   rƒ   rT   Úsemantic_model_configÚfcrE  Úembedr`  re  Ú	prior_netrU   r3   Ú
rotary_embÚrangeÚnum_hidden_layersÚlayersÚpost_netÚ	LayerNormr£  r™  Úheadr‡   s    `€r-   r=   zXcodec2Decoder.__init__9  s3  øø€ Ý‰Œ×ÒÑÔÐÝ”)˜FÔ.°Ô1MÔ1YÑYÐ[aÔ[mÑnÔnˆŒÝ”Y˜vÔ1°6Ô3EÐSTÐ^_Ð`Ñ`Ô`ˆŒ
ÝœÕ(:¸6Ñ(BÔ(BÕDVÐW]ÑD^ÔD^Ð'_Ñ`Ô`ˆŒØ#)Ô#=ˆÔ Ý0¸Ð?Ñ?Ô?ˆŒÝ”mØeÐeÐeÐeÅUÈ6ÔKcÑEdÔEdÐeÑeÔeñ
ô 
ˆŒõ œÕ'9¸&Ñ'AÔ'AÕCUÐV\ÑC]ÔC]Ð&^Ñ_Ô_ˆŒÝ”L Ô!3¸Ð>Ñ>Ô>ˆŒ	Ý$ VÑ,Ô,ˆŒ	ˆ	ˆ	r,   rˆ   rL   c                 ó  — |                       |¦  «        }|                     dd¦  «        }|                      |¦  «        }|                     dd¦  «        }| j        D ]} ||¦  «        }Œt	          j        | j        |j        ¬¦  «                             d¦  «        }|  	                    ||¦  «        }| j
        D ]} ||fd|i|¤Ž}Œ| j        D ]} ||¦  «        }Œ|                      |                      |¦  «        ¦  «        S )Nr   r   r~  r   rÄ   )rÊ  rk   rË  rÌ  r&   rV   rU   r4   r’   rÍ  rÐ  rÑ  rÓ  r£  )rI   rˆ   rY   Úlayerrp   rÄ   s         r-   ru   zXcodec2Decoder.forwardG  s&  € ØŸš Ñ.Ô.ˆØ%×/Ò/°°1Ñ5Ô5ˆØŸ
š
 =Ñ1Ô1ˆØ%×/Ò/°°1Ñ5Ô5ˆð ”^ð 	1ð 	1ˆEØ!˜E -Ñ0Ô0ˆMˆMõ ”| DÔ$<À]ÔEYÐZÑZÔZ×dÒdÐefÑgÔgˆØ"Ÿošo¨m¸\ÑJÔJÐØ”[ð 	dð 	dˆEØ!˜E -ÐcÐcÐEXÐcÐ\bÐcÐcˆMˆMð ”]ð 	1ð 	1ˆEØ!˜E -Ñ0Ô0ˆMˆMà�yŠy˜Ÿš =Ñ1Ô1Ñ2Ô2Ð2r,   r¶  r{   s   @r-   rÄ  rÄ  6  sp   ø€ € € € € Ø`Ð`ð-˜}ð -ð -ð -ð -ð -ð -ð3 U¤\ð 3ÀÄð 3ð 3ð 3ð 3ð 3ð 3ð 3ð 3r,   rÄ  c                   óH   ‡ — e Zd Zdefˆ fd„Zdej        dej        fd„Zˆ xZS )ÚXcodec2SemanticAdapterr6   c                 ó8  •— t          ¦   «                              ¦   «          t          j        |j        j        |j        j        ddd¬¦  «        | _        t          j        ¦   «         | _        t          j        |j        j        |j        j        ddd¬¦  «        | _	        t          j        ¦   «         | _
        t          j        |j        j        |j        j        ddd¬¦  «        | _        t          j        |j        j        |j        j        ddd¬¦  «        | _        d S )Nr   r   F)Úin_channelsÚout_channelsr  rC  r€   T)r  rC  r€   )r<   r=   rD   rE  rÉ  rT   rF  ÚReLUÚact1rH  Úact2Úconv3Úconv4r‡   s     €r-   r=   zXcodec2SemanticAdapter.__init__c  s	  ø€ Ý‰Œ×ÒÑÔÐÝ”YØÔ4Ô@ØÔ5ÔAØØØð
ñ 
ô 
ˆŒ
õ ”G‘I”IˆŒ	Ý”YØÔ(Ô4ØÔ(Ô4ØØØð
ñ 
ô 
ˆŒ
õ ”G‘I”IˆŒ	Ý”YØÔ(Ô4ØÔ(Ô4ØØØð
ñ 
ô 
ˆŒ
õ ”YØÔ4Ô@ØÔ5ÔAØØØð
ñ 
ô 
ˆŒ
ˆ
ˆ
r,   rˆ   rL   c                 ó  — |                       |¦  «        }|                      |¦  «        }|}|                      |¦  «        }|                      |¦  «        }|                      |¦  «        }||z   }|                      |¦  «        }|S rv   )rF  rÜ  rH  rÝ  rÞ  rß  rs  s      r-   ru   zXcodec2SemanticAdapter.forward„  s}   € ØŸ
š
 =Ñ1Ô1ˆØŸ	š	 -Ñ0Ô0ˆØ ˆØŸ
š
 =Ñ1Ô1ˆØŸ	š	 -Ñ0Ô0ˆØŸ
š
 =Ñ1Ô1ˆØ%¨Ñ0ˆØŸ
š
 =Ñ1Ô1ˆØÐr,   r‹   r{   s   @r-   r×  r×  b  sk   ø€ € € € € ð
˜}ð 
ð 
ð 
ð 
ð 
ð 
ðB	 U¤\ð 	°e´lð 	ð 	ð 	ð 	ð 	ð 	ð 	ð 	r,   r×  c                   óf   ‡ — e Zd ZU eed<   dZdZdZdZdgZ	dZ
dZdZdZdZdZdZeedœZˆ fd	„Zˆ xZS )
ÚXcodec2PreTrainedModelr6   Úxcodec2)r´  TNrÅ   Úinput_values)rˆ   Ú
attentionsc                 óÒ  •— t          ¦   «                              |¦  «         t          |t          ¦  «        r4t	          j        |j        ¦  «         t	          j        |j        ¦  «         d S t          |t          ¦  «        r5t          j
        |j        ¦  «        }t	          j        |j        |¦  «         d S t          |t          ¦  «        rt|                     |j        j        ¬¦  «        \  }}}t	          j        |j        |¦  «         t	          j        |j        |¦  «         t	          j        |j        |¦  «         d S t          |t(          ¦  «        rBt+          d|j        z  d|j        z  |j        ¦  «        }t	          j        |j        |¦  «         d S t          |t2          ¦  «        r<t+          |j        |j        |j        ¦  «        }t	          j        |j        |¦  «         d S d S )Nr~  rÿ   r  )r<   Ú_init_weightsrh   rñ   ÚinitÚzeros_rõ   rö   r™  r&   rž  r›  Úcopy_rŸ  ru  ry  rz  r4   r{  r|  r)  r  r  r  r  r  r  r  )rI   r¡   rŸ  rz  r{  r|  Úfilter_tensorrK   s          €r-   rç  z$Xcodec2PreTrainedModel._init_weights¤  s¯  ø€ Ý‰Œ×Ò˜fÑ%Ô%Ð%Ý�fÕ.Ñ/Ô/ð 	5ÝŒK˜œÑ%Ô%Ð%ÝŒK˜œÑ$Ô$Ð$Ð$Ð$Ý˜Õ 0Ñ1Ô1ð 	5ÝÔ& v¤|Ñ4Ô4ˆFÝŒJ�v”} fÑ-Ô-Ð-Ð-Ð-Ý˜Õ ?Ñ@Ô@ð 
	5Ø&,×&=Ò&=ÀVÄ]ÔEYÐ&=Ñ&ZÔ&ZÑ#ˆF�E˜8ÝŒJ�v”} fÑ-Ô-Ð-ÝŒJ�v”| UÑ+Ô+Ð+ÝŒJ�v”¨Ñ1Ô1Ð1Ð1Ð1Ý˜Õ 1Ñ2Ô2ð 	5Ý0°°v´|Ñ1CÀSÈ6Ì<ÑEWÐY_ÔYkÑlÔlˆMÝŒJ�v”} mÑ4Ô4Ð4Ð4Ð4Ý˜Õ 3Ñ4Ô4ð 	5Ý0°´ÀÔ@QÐSYÔSeÑfÔfˆMÝŒJ�v”} mÑ4Ô4Ð4Ð4Ð4ð	5ð 	5r,   )r"   r#   r$   r   r(   Úbase_model_prefixÚinput_modalitiesÚsupports_gradient_checkpointingÚ_no_split_modulesÚ_skip_keys_device_placementÚ_supports_flash_attnÚ_supports_sdpaÚ_supports_flex_attnÚ_supports_cache_classÚ_supports_attention_backendÚ_can_compile_fullgraphÚmain_input_namerã   Ú_can_record_outputsrç  rz   r{   s   @r-   râ  râ  �  s¢   ø€ € € € € € àÐÐÑØ!ÐØ!ÐØ&*Ð#ØÐØ#4Ð"5ÐØÐØ€NØÐØ ÐØ"&ÐØ!ÐØ$€Oà,Ø)ðð Ðð
5ð 5ð 5ð 5ð 5ð 5ð 5ð 5ð 5r,   râ  z!Xcodec2 neural audio codec model.)Úcustom_introc                   ó¼  ‡ — e Zd ZeZdefˆ fd„Zee	 	 	 ddej	        dej	        dej	        dz  dej	        dz  d	e
d
ee         deez  fd„¦   «         ¦   «         Zee	 	 ddej	        dz  dej	        dz  d
ee         deez  fd„¦   «         ¦   «         Zee	 	 	 ddej	        dej	        dej	        dz  dej	        dz  d	e
d
ee         deez  fd„¦   «         ¦   «         Zˆ xZS )ÚXcodec2Modelr6   c                 óâ  •— t          ¦   «                              |¦  «         |j        | _        t          j        |j        ¦  «        | _        t          |¦  «        | _        t          |¦  «        | _
        t          j        |j        |j        j        z   |j        |j        j        z   ¦  «        | _        t          |¦  «        | _        t#          |¦  «        | _        |                      ¦   «          d S rv   )r<   r=   r�  r   Úfrom_configrÉ  Úsemantic_encoderr×  Úsemantic_adapterrZ  Úacoustic_encoderrD   rƒ   rT   Ú
fc_encoderr¸  rº  rÄ  Úacoustic_decoderÚ	post_initr‡   s     €r-   r=   zXcodec2Model.__init__½  sÇ   ø€ Ý‰Œ×Ò˜Ñ Ô Ð à Ô+ˆŒÝ )Ô 5°fÔ6RÑ SÔ SˆÔÝ 6°vÑ >Ô >ˆÔÝ .¨vÑ 6Ô 6ˆÔÝœ)ØÔ Ô!=Ô!IÑIØÔ Ô!=Ô!IÑIñ
ô 
ˆŒõ *¨&Ñ1Ô1ˆŒÝ .¨vÑ 6Ô 6ˆÔà�ŠÑÔÐÐÐr,   NFrä  Úinput_featuresÚpadding_maskÚinput_features_maskÚoutput_latentsrY   rL   c                 ó@  — t          j        ¦   «         5  |                      ||¬¦  «        }ddd¦  «         n# 1 swxY w Y   |j                             dd¦  «        }|                      |¦  «        }|                      |¦  «        }	t          j        ||	gd¬¦  «        }
|                      |
                     dd¦  «        ¦  «        }
|  	                    |
¦  «        \  }}|                     dd¦  «        }|                     dd¦  «        }d}|�y| 
                    dd¬¦  «        }|| j        z  }t          j        |j        d         |j        ¬	¦  «                             dd¦  «        }||k                          |j        ¦  «        }t%          ||r|nd|¬
¦  «        S )a  
        input_values (`torch.Tensor` of shape `(batch_size, 1, sequence_length)`):
            Input audio waveform.
        input_features (`torch.Tensor` of shape `(batch_size, mel_bins, time_steps)`):
            Input audio mel spectrogram for semantic encoding.
        padding_mask (`torch.Tensor` of shape `(batch_size, 1, sequence_length)`):
            Padding mask used to pad `input_values`.
        input_features_mask (`torch.Tensor` of shape `(batch_size, time_steps)`, *optional*):
            Attention mask for the spectrogram input to the semantic encoder. `1` for valid frames, `0` for padding.
        output_latents (`bool`, *optional*, defaults to `False`):
            Whether to return the continuous latent representation from the quantizer.
        )r¥   Nr   r   re   r^   T)r[   rÚ   r~  )r   r    r!   )r&   ry   rþ  Úlast_hidden_staterk   rÿ  r   rl   r  rº  r  r�  rV   rg   r4   rÇ   rX   rR   r/   )rI   rä  r  r  r  r  rY   Úsemantic_outputÚsemantic_hidden_statesÚacoustic_hidden_statesrˆ   r    r   r!   Úaudio_lengthÚtoken_lengthÚidxs                    r-   ÚencodezXcodec2Model.encodeÍ  sû  € õ2 Œ]‰_Œ_ð 	hð 	hØ"×3Ò3°NÐSfÐ3ÑgÔgˆOð	hð 	hð 	hñ 	hô 	hð 	hð 	hð 	hð 	hð 	hð 	høøøð 	hð 	hð 	hð 	hà!0Ô!B×!LÒ!LÈQÐPQÑ!RÔ!RÐØ!%×!6Ò!6Ð7MÑ!NÔ!NÐð "&×!6Ò!6°|Ñ!DÔ!DÐÝœ	Ð#9Ð;QÐ"RÐXYÐZÑZÔZˆØŸš¨×(?Ò(?ÀÀ1Ñ(EÔ(EÑFÔFˆð  $Ÿ~š~¨mÑ<Ô<Ñˆ�Ø×#Ò# A qÑ)Ô)ˆØ!×+Ò+¨A¨qÑ1Ô1ˆð  ÐØÐ#Ø'×+Ò+°¸DÐ+ÑAÔAˆLØ'¨4¬?Ñ:ˆLÝ”,˜{Ô0°Ô4¸\Ô=PÐQÑQÔQ×VÒVÐWXÐZ\Ñ]Ô]ˆCØ # lÒ 2×6Ò6°|Ô7IÑJÔJÐå#Ø#Ø-Ð7�G�G°4Ø-ð
ñ 
ô 
ð 	
s   ”8¸<¿<r   r    c                 óò   — |€|€t          d¦  «        ‚|�/| j                             |                     dd¦  «        ¦  «        }n|                     dd¦  «        } | j        |fi |¤Ž}t          |¬¦  «        S )a3  
        audio_codes (`torch.LongTensor`  of shape `(batch_size, 1, codes_length)`):
            Discrete code indices computed using `model.encode`.
        latents (torch.Tensor of shape `(batch_size, dimension, time_steps)`, *optional*):
            Quantized continuous representation of input.
        Nz3Either `latents` or `audio_codes` must be provided.r   r   )r   )r  rº  rÀ  rk   r  r1   )rI   r   r    rY   Úrecon_audios        r-   ÚdecodezXcodec2Model.decode  sŠ   € ð ˆ?˜{Ð2ÝÐRÑSÔSÐSàÐ"Ø”n×/Ò/°×0EÒ0EÀaÈÑ0KÔ0KÑLÔLˆGˆGà×'Ò'¨¨1Ñ-Ô-ˆGà+�dÔ+¨GÐ>Ð>°vÐ>Ð>ˆÝ#°Ð=Ñ=Ô=Ð=r,   c                 óè   — |j         d         }|                      ||||dd¬¦  «        } | j        d	|j        ddœ|¤Žd         dd|…f         }	t	          |	|j        |r|j        nd|j        ¬¦  «        S )
a  
        input_values (`torch.Tensor` of shape `(batch_size, 1, sequence_length)`):
            Input audio waveform.
        input_features (`torch.Tensor` of shape `(batch_size, mel_bins, time_steps)`):
            Input audio mel spectrogram for semantic encoding.
        padding_mask (`torch.Tensor` of shape `(batch_size, 1, sequence_length)`):
            Padding mask used to pad `input_values`.
        input_features_mask (`torch.Tensor` of shape `(batch_size, time_steps)`, *optional*):
            Attention mask for the spectrogram input to the semantic encoder. `1` for valid frames, `0` for padding.
        output_latents (`bool`, *optional*, defaults to `False`):
            Whether to return the continuous latent representation from the quantizer.

        Examples:

        ```python
        >>> from datasets import load_dataset
        >>> from transformers import AutoFeatureExtractor, Xcodec2Model

        >>> dataset = load_dataset("hf-internal-testing/librispeech_asr_dummy", "clean", split="validation")
        >>> audio = dataset["train"]["audio"][0]["array"]

        >>> model_id = "HKUSTAudio/xcodec2-hf"
        >>> model = Xcodec2Model.from_pretrained(model_id)
        >>> feature_extractor = AutoFeatureExtractor.from_pretrained(model_id)

        >>> inputs = feature_extractor(audio=audio, sampling_rate=feature_extractor.sampling_rate, return_tensors="pt")

        >>> outputs = model(**inputs)
        >>> audio_codes = outputs.audio_codes
        >>> audio_values = outputs.audio_values
        ```r^   T)r  r  r  r  Úreturn_dict)r    r  r   .N)r   r   r    r!   r+   )rg   r  r  r    r   r   r!   )
rI   rä  r  r  r  r  rY   ÚlengthÚencoder_outputsr   s
             r-   ru   zXcodec2Model.forward  s¯   € ðV Ô# BÔ'ˆàŸ+š+ØØ)Ø%Ø 3ØØð &ñ 
ô 
ˆð #�t”{Ð_¨?Ô+BÐPTÐ_Ð_ÐX^Ð_Ð_Ð`aÔbÐcfÐhoÐioÐhoÐcoÔpˆåØ%Ø'Ô3Ø/=ÐG�OÔ+Ð+À4Ø,Ô=ð	
ñ 
ô 
ð 	
r,   )NNF)NN)r"   r#   r$   r   Úconfig_classr=   r   r   r&   r*   rï   r   r   rx   r/   r  r1   r  r   ru   rz   r{   s   @r-   rû  rû  ¹  sû  ø€ € € € € à €Lð˜}ð ð ð ð ð ð ð  Øð
 -1Ø37Ø$ð2
ð 2
à”lð2
ð œð2
ð ”l TÑ)ð	2
ð
 #œ\¨DÑ0ð2
ð ð2
ð Ð+Ô,ð2
ð 
Ð%Ñ	%ð2
ð 2
ð 2
ñ Ôñ „^ð2
ðh Øð ,0Ø'+ð>ð >à”\ DÑ(ð>ð ” Ñ$ð>ð Ð+Ô,ð	>ð
 
Ð%Ñ	%ð>ð >ð >ñ Ôñ „^ð>ð. Øð
 -1Ø37Ø$ð:
ð :
à”lð:
ð œð:
ð ”l TÑ)ð	:
ð
 #œ\¨DÑ0ð:
ð ð:
ð Ð+Ô,ð:
ð 
�Ñ	ð:
ð :
ð :
ñ Ôñ „^ð:
ð :
ð :
ð :
ð :
r,   rû  )r   )r    )Pr   Úcollections.abcr   Údataclassesr   Únumpyr‚  r&   Útorch.nnrD   Útorch.nn.functionalr¯   r!  r   Ú r   rè  Úactivationsr   Úcache_utilsr	   Úintegrationsr
   r   Úmodeling_layersr   Úmodeling_rope_utilsr   r   Úmodeling_utilsr   r   Úprocessing_utilsr   Úutilsr   r   r   r   Úutils.deprecationr   Úutils.genericr   Úautor   Úconfiguration_xcodec2r   r   r/   r1   ÚModuler3   r}   r�   r˜   r*   rÏ   rŸ   rW   r·   r¹   rÒ   rã   rñ   r  r  r)  r/  r>  rN  rZ  re  ru  r™  r¸  rÄ  r×  râ  rû  Ú__all__r+   r,   r-   ú<module>r-     sX  ðð( €€€Ø $Ð $Ð $Ð $Ð $Ð $Ø !Ð !Ð !Ð !Ð !Ð !à Ð Ð Ð Ø €€€Ø Ð Ð Ð Ð Ð Ø Ð Ð Ð Ð Ð Ð Ð Ð Ø Ð Ð Ð Ð Ð à &Ð &Ð &Ð &Ð &Ð &Ø !Ð !Ð !Ð !Ð !Ð !Ø  Ð  Ð  Ð  Ð  Ð  Ø LÐ LÐ LÐ LÐ LÐ LÐ LÐ LØ 9Ð 9Ð 9Ð 9Ð 9Ð 9Ø KÐ KÐ KÐ KÐ KÐ KÐ KÐ KØ FÐ FÐ FÐ FÐ FÐ FÐ FÐ FØ &Ð &Ð &Ð &Ð &Ð &Ø VÐ VÐ VÐ VÐ VÐ VÐ VÐ VÐ VÐ VÐ VÐ VØ 0Ð 0Ð 0Ð 0Ð 0Ð 0Ø +Ð +Ð +Ð +Ð +Ð +Ø Ð Ð Ð Ð Ð Ø 0Ð 0Ð 0Ð 0Ð 0Ð 0ð Ø
ð1ð 1ð 1ð 1ð 1�Kñ 1ô 1ñ „ñ „ð1ð( Ø
ð1ð 1ð 1ð 1ð 1˜;ñ 1ô 1ñ „ñ „ð1ð" Ø
ð2ð 2ð 2ð 2ð 2˜;ñ 2ô 2ñ „ñ „ð2ð6<ð 6<ð 6<ð 6<ð 6<˜RœYñ 6<ô 6<ð 6<ðrð ð ð ð �”ñ ô ð ð(ð (ð (ð ÐÐ-Ñ.Ô.ðð ð ñ /Ô.ðð2	U˜Uœ\ð 	U°#ð 	U¸%¼,ð 	Uð 	Uð 	Uð 	Uð& ð%ð %ØŒIð%àŒ<ð%ð 
Œð%ð Œ<ð	%ð
 ”L 4Ñ'ð%ð ð%ð ð%ð Ð'Ô(ð%ð %ð %ð %ð2 ÐÐ)Ñ*Ô*ðC)ð C)ð C)ð C)ð C)�r”yñ C)ô C)ñ +Ô*ðC)ðL Ð˜YÑ'Ô'ðJð Jð Jð Jð J�R”Yñ Jô Jñ (Ô'ðJð((ð (ð (ð (ð (Ð4ñ (ô (ð (ðV&ð &ð &ð &ð &�r”yñ &ô &ð &ðR+5ð +5ð +5ð\ð ð ð ð ˜"œ)ñ ô ð ðDð ð ð ð ˜œ	ñ ô ð ð6ð ð ð ð  R¤Yñ ô ð ð0!ð !ð !ð !ð !˜"œ)ñ !ô !ð !ðHð ð ð ð ˜"œ)ñ ô ð ð.ð ð ð ð �R”Yñ ô ð ð@:ð :ð :ð :ð :˜œñ :ô :ð :ð0P1ð P1ð P1ð P1ð P1 b¤iñ P1ô P1ð P1ðf3"ð 3"ð 3"ð 3"ð 3"�r”yñ 3"ô 3"ð 3"ðl&ð &ð &ð &ð &�r”yñ &ô &ð &ð,)3ð )3ð )3ð )3ð )3�R”Yñ )3ô )3ð )3ðX+ð +ð +ð +ð +˜RœYñ +ô +ð +ð\ ð%5ð %5ð %5ð %5ð %5˜_ñ %5ô %5ñ „ð%5ðP €Ð@ÐAÑAÔAð^
ð ^
ð ^
ð ^
ð ^
Ð)ñ ^
ô ^
ñ BÔAð^
ðB Ð3Ð
4€€€r,   