Ë
    ÛÓ^jJo  ã                   ó  — d dl mZ d dlmZ d dl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mZ dd	lmZ dd
lmZmZmZmZ ddlmZmZ ddl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&m'Z' ddl(m)Z) ddl*m+Z+m,Z,m-Z- ddl.m/Z/m0Z0 ddl1m2Z2 ddl3m4Z4  ed«       G d„ dejj                  «      «       Z6 G d„ dejj                  «      Z7d„ Z8 ed«      d?d„«       Z9dejt                  d e;d!ejt                  fd"„Z<	 d@d#ejj                  d$ejt                  d%ejt                  d&ejt                  d'ejt                  dz  d(e=d)e=d*e)e+   fd+„Z> ee9«       G d,„ d-ejj                  «      «       Z? G d.„ d/ejj                  «      Z@ G d0„ d1ejj                  «      ZAe G d2„ d3ejj                  «      «       ZB G d4„ d5ejj                  «      ZC G d6„ d7e«      ZDe, G d8„ d9e'«      «       ZEe, G d:„ d;eE«      «       ZFe, G d<„ d=eEe«      «       ZGg d>¢ZHy)Aé    )ÚCallable)ÚOptionalN)Únné   )Úinitialization)ÚACT2FN)ÚCacheÚDynamicCache)ÚGenerationMixin)Úuse_experts_implementationÚuse_kernel_forward_from_hubÚuse_kernel_func_from_hubÚuse_kernelized_func)Úcreate_causal_maskÚ!create_sliding_window_causal_mask)ÚFlashAttentionKwargs)ÚGradientCheckpointingLayer)ÚBaseModelOutputWithPastÚCausalLMOutputWithPast)ÚROPE_INIT_FUNCTIONSÚdynamic_rope_update)ÚALL_ATTENTION_FUNCTIONSÚPreTrainedModel)ÚUnpack)ÚTransformersKwargsÚauto_docstringÚcan_return_tuple)Úmaybe_autocastÚmerge_with_config_defaults)Úcapture_outputsé   )ÚDots1ConfigÚRMSNormc                   óh   ‡ — 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Dots1RMSNormÚepsÚreturnNc                 óŠ   •— t         ‰| �  «        t        j                  t	        j
                  |«      «      | _        || _        y)z;
        Dots1RMSNorm is equivalent to T5LayerNorm
        N)ÚsuperÚ__init__r   Ú	ParameterÚtorchÚonesÚweightÚvariance_epsilon)ÚselfÚhidden_sizer&   Ú	__class__s      €ús/var/www/ramen.bs-engineer-server.com/venv/lib/python3.12/site-packages/transformers/models/dots1/modeling_dots1.pyr*   zDots1RMSNorm.__init__4   s1   ø€ ô 	‰ÑÔÜ—l‘l¤5§:¡:¨kÓ#:Ó;ˆŒØ #ˆÕó    Úhidden_statesc                 ó"  — |j                   }|j                  t        j                  «      }|j	                  d«      j                  dd¬«      }|t        j                  || j                  z   «      z  }| j                  |j                  |«      z  S )Né   éÿÿÿÿT)Úkeepdim)	ÚdtypeÚtor,   Úfloat32ÚpowÚmeanÚrsqrtr/   r.   )r0   r5   Úinput_dtypeÚvariances       r3   ÚforwardzDots1RMSNorm.forward<   sy   € Ø#×)Ñ)ˆØ%×(Ñ(¬¯©Ó7ˆØ ×$Ñ$ QÓ'×,Ñ,¨R¸Ð,Ó>ˆØ%¬¯©°H¸t×?TÑ?TÑ4TÓ(UÑUˆØ�{‰{˜]×-Ñ-¨kÓ:Ñ:Ð:r4   c                 ó^   — t        | j                  j                  «      › d| j                  › �S )Nz, eps=)Útupler.   Úshaper/   )r0   s    r3   Ú
extra_reprzDots1RMSNorm.extra_reprC   s*   € Ü˜Ÿ™×)Ñ)Ó*Ð+¨6°$×2GÑ2GÐ1HÐIÐIr4   )g�íµ ÷Æ°>)
Ú__name__Ú
__module__Ú__qualname__Úfloatr*   r,   ÚTensorrB   rF   Ú__classcell__©r2   s   @r3   r%   r%   2   s7   ø„ ñ$¨ð $¸$õ $ð; U§\¡\ð ;°e·l±ló ;öJr4   r%   c                   óÄ   ‡ — e Zd ZU ej                  ed<   ddefˆ fd„Ze	 	 	 ddedz  de	d   de
dz  ded	ef   fd
„«       Z ej                  «       ed„ «       «       Zˆ xZS )ÚDots1RotaryEmbeddingÚinv_freqNÚconfigc                 ó®  •— t         ‰| �  «        |j                  | _        |j                  | _        || _        | j
                  j                  d   | _        | j                  }| j                  dk7  rt        | j                     } || j
                  |«      \  }| _
        | j                  d|d¬«       | j                  d|j                  «       d¬«       y )NÚ	rope_typeÚdefaultrP   F)Ú
persistentÚoriginal_inv_freq)r)   r*   Úmax_position_embeddingsÚmax_seq_len_cachedÚoriginal_max_seq_lenrQ   Úrope_parametersrS   Úcompute_default_rope_parametersr   Úattention_scalingÚregister_bufferÚclone)r0   rQ   ÚdeviceÚrope_init_fnrP   r2   s        €r3   r*   zDots1RotaryEmbedding.__init__J   s¶   ø€ Ü‰ÑÔØ"(×"@Ñ"@ˆÔØ$*×$BÑ$BˆÔ!àˆŒàŸ™×4Ñ4°[ÑAˆŒØ!%×!EÑ!EˆØ�>‰>˜YÒ&Ü.¨t¯~©~Ñ>ˆLÙ+7¸¿¹ÀVÓ+LÑ(ˆ�$Ô(à×Ñ˜Z¨¸eÐÔDØ×ÑÐ0°(·.±.Ó2BÈuÐÕUr4   r_   ztorch.deviceÚseq_lenr'   ztorch.Tensorc                 ó  — | j                   d   }t        | dd«      xs | j                  | j                  z  }d}d|t	        j
                  d|dt        j                  ¬«      j                  |t        j                  ¬«      |z  z  z  }||fS )	a¨  
        Computes the inverse frequencies according to the original RoPE implementation
        Args:
            config ([`~transformers.PreTrainedConfig`]):
                The model configuration.
            device (`torch.device`):
                The device to use for initialization of the inverse frequencies.
            seq_len (`int`, *optional*):
                The current sequence length. Unused for this type of RoPE.
        Returns:
            Tuple of (`torch.Tensor`, `float`), containing the inverse frequencies for the RoPE embeddings and the
            post-processing scaling factor applied to the computed cos/sin (unused in this type of RoPE).
        Ú
rope_thetaÚhead_dimNg      ð?r   r7   ©r:   )r_   r:   )	rZ   Úgetattrr1   Únum_attention_headsr,   ÚarangeÚint64r;   rJ   )rQ   r_   ra   ÚbaseÚdimÚattention_factorrP   s          r3   r[   z4Dots1RotaryEmbedding.compute_default_rope_parametersZ   s�   € ð& ×%Ñ% lÑ3ˆÜ�f˜j¨$Ó/Òc°6×3EÑ3EÈ×IcÑIcÑ3cˆàÐð Ø”U—\‘\ ! S¨!´5·;±;Ô?×BÑBÈ&ÔX]×XcÑXcÐBÓdÐgjÑjÑkñ
ˆð Ð)Ð)Ð)r4   c                 óN  — | j                   d d d …d f   j                  «       j                  |j                  d   dd«      j	                  |j
                  «      }|d d …d d d …f   j                  «       }t        |j
                  j                  t        «      r/|j
                  j                  dk7  r|j
                  j                  nd}t        |d¬«      5  |j                  «       |j                  «       z  j                  dd«      }t        j                  ||fd¬	«      }|j                  «       | j                  z  }|j                  «       | j                  z  }	d d d «       j	                  |j                   ¬
«      	j	                  |j                   ¬
«      fS # 1 sw Y   ŒAxY w)Nr   r8   r!   ÚmpsÚcpuF)Údevice_typeÚenabledr7   ©rk   re   )rP   rJ   ÚexpandrE   r;   r_   Ú
isinstanceÚtypeÚstrr   Ú	transposer,   ÚcatÚcosr\   Úsinr:   )
r0   ÚxÚposition_idsÚinv_freq_expandedÚposition_ids_expandedrp   ÚfreqsÚembry   rz   s
             r3   rB   zDots1RotaryEmbedding.forwardx   sR  € ð !ŸM™M¨$²°4¨-Ñ8×>Ñ>Ó@×GÑGÈ×HZÑHZÐ[\ÑH]Ð_aÐcdÓe×hÑhÐij×iqÑiqÓrÐØ ,ªQ°²a¨ZÑ 8× >Ñ >Ó @Ðä'1°!·(±(·-±-ÄÔ'EÈ!Ï(É(Ï-É-Ð[`ÒJ`�a—h‘h—m’mÐfkˆÜ¨¸UÔCñ 	5Ø&×,Ñ,Ó.Ð1F×1LÑ1LÓ1NÑN×YÑYÐZ[Ð]^Ó_ˆEÜ—)‘)˜U E˜N°Ô3ˆCØ—'‘'“)˜d×4Ñ4Ñ4ˆCØ—'‘'“)˜d×4Ñ4Ñ4ˆC÷		5ð �v‰v˜AŸG™GˆvÓ$ c§f¡f°1·7±7 fÓ&;Ð;Ð;÷	5ð 	5ús   ÃBFÆF$©N)NNN)rG   rH   rI   r,   rK   Ú__annotations__r"   r*   Ústaticmethodr   ÚintrD   rJ   r[   Úno_gradr   rB   rL   rM   s   @r3   rO   rO   G   s˜   ø… Ø�l‰lÓñV˜{õ Vð  à%)Ø+/Ø"ñ*Ø˜dÑ"ð*à˜Ñ(ð*ð �t‘ð*ð 
ˆ~˜uÐ$Ñ	%ò	*ó ð*ð: €U‡]�]ƒ_Øñ<ó ó ô<r4   rO   c                 óš   — | 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..Nr8   r7   rr   )rE   r,   rx   )r{   Úx1Úx2s      r3   Úrotate_halfr‰   ˆ   sZ   € à	
ˆ3Ð"�!—'‘'˜"‘+ Ñ"Ð"Ð"Ñ	#€BØ	
ˆ3�—‘˜‘˜qÑ Ñ"Ð"Ñ	#€BÜ�9‰9�r�c˜2�Y BÔ'Ð'r4   Úrotary_pos_embc                 óž   — |j                  |«      }|j                  |«      }| |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kry   rz   Úunsqueeze_dimÚq_embedÚk_embeds          r3   Úapply_rotary_pos_embr’   �   sY   € ð& �-‰-˜Ó
&€CØ
�-‰-˜Ó
&€CØ�3‰wœ; q›>¨CÑ/Ñ0€GØ�3‰wœ; q›>¨CÑ/Ñ0€GØ�GÐÐr4   r5   Ún_repr'   c                 óª   — | j                   \  }}}}|dk(  r| S | dd…dd…ddd…dd…f   j                  |||||«      } | j                  |||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)rE   rs   Úreshape)r5   r“   ÚbatchÚnum_key_value_headsÚslenrd   s         r3   Ú	repeat_kvr™   ©   so   € ð
 2?×1DÑ1DÑ.€EÐ  hØ�‚zØÐØ!¢!¢Q¨ªa²Ð"2Ñ3×:Ñ:¸5ÐBUÐW\Ð^bÐdlÓm€MØ× Ñ  Ð(;¸eÑ(CÀTÈ8ÓTÐTr4   ÚmoduleÚqueryÚkeyÚvalueÚattention_maskÚscalingÚdropoutÚkwargsc                 ó  — t        || j                  «      }t        || j                  «      }	t        j                  ||j	                  dd«      «      |z  }
|�|
|z   }
t
        j                  j                  |
dt        j                  ¬«      j                  |j                  «      }
t
        j                  j                  |
|| j                  ¬«      }
t        j                  |
|	«      }|j	                  dd«      j                  «       }||
fS )Nr7   r   r8   )rk   r:   )ÚpÚtrainingr!   )r™   Únum_key_value_groupsr,   Úmatmulrw   r   Ú
functionalÚsoftmaxr<   r;   r:   r    r¤   Ú
contiguous)rš   r›   rœ   r�   rž   rŸ   r    r¡   Ú
key_statesÚvalue_statesÚattn_weightsÚattn_outputs               r3   Úeager_attention_forwardr®   µ   sà   € ô ˜3 × ;Ñ ;Ó<€JÜ˜U F×$?Ñ$?Ó@€Lä—<‘<  z×';Ñ';¸A¸qÓ'AÓBÀWÑL€LØÐ!Ø# nÑ4ˆä—=‘=×(Ñ(¨¸2ÄUÇ]Á]Ð(ÓS×VÑVÐW\×WbÑWbÓc€LÜ—=‘=×(Ñ(¨¸È6Ï?É?Ð(Ó[€LÜ—,‘,˜|¨\Ó:€KØ×'Ñ'¨¨1Ó-×8Ñ8Ó:€Kà˜Ð$Ð$r4   c                   óô   ‡ — 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ej                  dz  d	e
dz  d
ee   de	ej                  ej                  dz  f   fd„Zˆ xZS )ÚDots1Attentionz=Multi-headed attention from 'Attention Is All You Need' paperrQ   Ú	layer_idxc                 ó�  •— t         ‰| �  «        t        |d«      r|j                  |   nd | _        || _        || _        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$                  ¬«      | _        t!        j"                  |j                  | j                  z  |j                  |j$                  ¬«      | _        t/        | j                  |j0                  ¬«      | _        t/        | j                  |j0                  ¬«      | _        | j                  dk(  r|j6                  | _        y d | _        y )NÚlayer_typesrd   g      à¿T©Úbias©r&   Úsliding_attention)r)   r*   Úhasattrr³   Ú
layer_typerQ   r±   rf   r1   rg   rd   r—   r¥   rŸ   Úattention_dropoutÚ	is_causalr   ÚLinearÚattention_biasÚq_projÚk_projÚv_projÚo_projr%   Úrms_norm_epsÚq_normÚk_normÚsliding_window©r0   rQ   r±   r2   s      €r3   r*   zDots1Attention.__init__Ò   s¿  ø€ Ü‰ÑÔÜ;BÀ6È=Ô;Y˜&×,Ñ,¨YÒ7Ð_cˆŒØˆŒØ"ˆŒÜ ¨
°F×4FÑ4FÈ&×JdÑJdÑ4dÓeˆŒØ$*×$>Ñ$>À&×B\ÑB\Ñ$\ˆÔ!Ø—}‘} dÑ*ˆŒØ!'×!9Ñ!9ˆÔØˆŒä—i‘iØ×Ñ × :Ñ :¸T¿]¹]Ñ JÐQW×QfÑQfô
ˆŒô —i‘iØ×Ñ × :Ñ :¸T¿]¹]Ñ JÐQW×QfÑQfô
ˆŒô —i‘iØ×Ñ × :Ñ :¸T¿]¹]Ñ JÐQW×QfÑQfô
ˆŒô —i‘iØ×&Ñ&¨¯©Ñ6¸×8JÑ8JÐQW×QfÑQfô
ˆŒô # 4§=¡=°f×6IÑ6IÔJˆŒÜ" 4§=¡=°f×6IÑ6IÔJˆŒØ7;·±ÐJ]Ò7]˜f×3Ñ3ˆÕÐcgˆÕr4   Nr5   Úposition_embeddingsrž   Úpast_key_valuesr¡   r'   c                 ó\  — |j                   d d }g |¢d‘| j                  ‘­}| j                  | j                  |«      j	                  |«      «      j                  dd«      }| j                  | j                  |«      j	                  |«      «      j                  dd«      }	| j                  |«      j	                  |«      j                  dd«      }
|\  }}t        ||	||«      \  }}	|� |j                  |	|
| j                  «      \  }	}
t        j                  | j                  j                  t         «      } || ||	|
|f| j"                  sdn| j$                  | j&                  | j(                  dœ|¤Ž\  }} |j*                  g |¢d‘­Ž j-                  «       }| j/                  |«      }||fS )Nr8   r!   r7   ç        )r    rŸ   rÅ   )rE   rd   rÃ   r¾   Úviewrw   rÄ   r¿   rÀ   r’   Úupdater±   r   Úget_interfacerQ   Ú_attn_implementationr®   r¤   rº   rŸ   rÅ   r•   r©   rÁ   )r0   r5   rÇ   rž   rÈ   r¡   Úinput_shapeÚhidden_shapeÚquery_statesrª   r«   ry   rz   Úattention_interfacer­   r¬   s                   r3   rB   zDots1Attention.forwardí   s®  € ð $×)Ñ)¨#¨2Ð.ˆØ8˜Ð8 bÐ8¨$¯-©-Ñ8ˆà—{‘{ 4§;¡;¨}Ó#=×#BÑ#BÀ<Ó#PÓQ×[Ñ[Ð\]Ð_`ÓaˆØ—[‘[ §¡¨]Ó!;×!@Ñ!@ÀÓ!NÓO×YÑYÐZ[Ð]^Ó_ˆ
Ø—{‘{ =Ó1×6Ñ6°|ÓD×NÑNÈqÐRSÓTˆà&‰ˆˆSÜ#7¸ÀjÐRUÐWZÓ#[Ñ ˆ�jàÐ&Ø'6×'=Ñ'=¸jÈ,ÐX\×XfÑXfÓ'gÑ$ˆJ˜ä(?×(MÑ(MØ�K‰K×,Ñ,Ô.Eó)
Ðñ %8ØØØØØð
%
ð  $Ÿ}š}‘C°$×2HÑ2HØ—L‘LØ×.Ñ.ñ
%
ð ñ
%
Ñ!ˆ�\ð *�k×)Ñ)Ð;¨;Ð;¸Ò;×FÑFÓHˆØ—k‘k +Ó.ˆØ˜LÐ(Ð(r4   r�   )rG   rH   rI   Ú__doc__r"   r„   r*   r,   rK   rD   r	   r   r   rB   rL   rM   s   @r3   r°   r°   Î   s¡   ø„ áGðh˜{ð h°sõ hð@ )-ñ')à—|‘|ð')ð # 5§<¡<°·±Ð#=Ñ>ð')ð Ÿ™ tÑ+ð	')ð
  ™ð')ð Ð-Ñ.ð')ð 
ˆu�|‰|˜UŸ\™\¨DÑ0Ð0Ñ	1÷')r4   r°   c                   ó&   ‡ — e Zd Zdˆ fd„	Zd„ Zˆ xZS )ÚDots1MLPc                 óÒ  •— t         ‰| �  «        || _        |j                  | _        |€|j                  n|| _        t        j                  | j                  | j                  d¬«      | _        t        j                  | j                  | j                  d¬«      | _        t        j                  | j                  | j                  d¬«      | _	        t        |j                     | _        y ©NFr´   )r)   r*   rQ   r1   Úintermediate_sizer   r¼   Ú	gate_projÚup_projÚ	down_projr   Ú
hidden_actÚact_fn)r0   rQ   rØ   r2   s      €r3   r*   zDots1MLP.__init__  s®   ø€ Ü‰ÑÔØˆŒØ!×-Ñ-ˆÔØ=NÐ=V ×!9Ò!9Ð\mˆÔÜŸ™ 4×#3Ñ#3°T×5KÑ5KÐRWÔXˆŒÜ—y‘y ×!1Ñ!1°4×3IÑ3IÐPUÔVˆŒÜŸ™ 4×#9Ñ#9¸4×;KÑ;KÐRWÔXˆŒÜ˜V×.Ñ.Ñ/ˆ�r4   c                 óˆ   — | j                  | j                  | j                  |«      «      | j                  |«      z  «      }|S r�   )rÛ   rÝ   rÙ   rÚ   )r0   r{   rÛ   s      r3   rB   zDots1MLP.forward"  s6   € Ø—N‘N 4§;¡;¨t¯~©~¸aÓ/@Ó#AÀDÇLÁLÐQRÃOÑ#SÓTˆ	ØÐr4   r�   ©rG   rH   rI   r*   rB   rL   rM   s   @r3   rÕ   rÕ     s   ø„ õ0ör4   rÕ   c                   ó$   ‡ — e Zd Zˆ fd„Zd„ Zˆ xZS )ÚDots1TopkRouterc                 óò  •— t         ‰| �  «        |j                  | _        |j                  | _        |j                  | _        t        j                  t        j                  | j
                  | j                  «      «      | _        |j                  | _        |j                  | _        |j                   | _        |j"                  | _        | j%                  dt        j                  | j
                  «      «       y )NÚe_score_correction_bias)r)   r*   Únum_experts_per_tokÚtop_kÚnum_local_expertsÚnum_expertsr1   Ú
hidden_dimr   r+   r,   Úzerosr.   Úrouted_scaling_factorÚn_groupÚ	num_groupÚ
topk_groupÚnorm_topk_probr]   ©r0   rQ   r2   s     €r3   r*   zDots1TopkRouter.__init__(  s¯   ø€ Ü‰ÑÔØ×/Ñ/ˆŒ
Ø!×3Ñ3ˆÔØ ×,Ñ,ˆŒÜ—l‘l¤5§;¡;¨t×/?Ñ/?ÀÇÁÓ#QÓRˆŒØ%+×%AÑ%AˆÔ"ØŸ™ˆŒØ ×+Ñ+ˆŒØ$×3Ñ3ˆÔØ×ÑÐ6¼¿¹ÀD×DTÑDTÓ8UÕVr4   c                 ól  — |j                  d| j                  «      }t        j                  |j	                  t
        j                  «      | j                  j	                  t
        j                  «      «      }|j                  «       }|| j                  z   }|j                  d| j                  | j                  | j                  z  «      j                  dd¬«      d   j                  d¬«      }t        j                  || j                  dd¬«      d   }t        j                  |«      }|j!                  d|d«       |j#                  d«      j%                  d| j                  | j                  | j                  z  «      j'                  d| j                  «      }|j)                  |j+                  «        t-        d«      «      }t        j                  || j.                  dd¬«      d   }	|j1                  d|	«      }
| j2                  r|
j                  dd	¬
«      dz   }|
|z  }
|
| j4                  z  }
||
|	fS )Nr8   r7   rr   r   F)rŽ   rk   Úsortedr!   z-infT)rk   r9   g#B’¡œÇ;)rË   rè   ÚFÚlinearru   r,   r<   r.   Úsigmoidrã   rì   rç   ÚtopkÚsumrí   Ú
zeros_likeÚscatter_rŒ   rs   r•   Úmasked_fillÚboolrJ   rå   Úgatherrî   rê   )r0   r5   Úrouter_logitsÚscoresÚscores_for_choiceÚgroup_scoresÚ	group_idxÚ
group_maskÚ
score_maskÚtopk_indicesÚtopk_weightsÚdenominators               r3   rB   zDots1TopkRouter.forward4  sÛ  € Ø%×*Ñ*¨2¨t¯©Ó?ˆÜŸ™ ×!3Ñ!3´E·M±MÓ!BÀDÇKÁK×DTÑDTÔUZ×UbÑUbÓDcÓdˆØ×&Ñ&Ó(ˆØ" T×%AÑ%AÑAÐà×"Ñ" 2 t§~¡~°t×7GÑ7GÈ4Ï>É>Ñ7YÓZß‰T�!˜ˆT‹_˜Qñ ç‰S�RˆS‹[ð 	ô
 —J‘J˜|¨t¯©ÀBÈuÔUÐVWÑXˆ	Ü×%Ñ% lÓ3ˆ
Ø×Ñ˜A˜y¨!Ô,à× Ñ  Ó$ß‰V�B˜Ÿ™¨×(8Ñ(8¸D¿N¹NÑ(JÓKß‰W�R˜×)Ñ)Ó*ð 	ð
 .×9Ñ9¸:¿?¹?Ó;LÐ:LÌeÐTZËmÓ\ÐÜ—z‘zÐ"3°t·z±zÀrÐRWÔXÐYZÑ[ˆØ—}‘} Q¨Ó5ˆØ×ÒØ&×*Ñ*¨r¸4Ð*Ó@À5ÑHˆKØ˜KÑ'ˆLØ# d×&@Ñ&@Ñ@ˆØ˜l¨LÐ8Ð8r4   rß   rM   s   @r3   rá   rá   '  s   ø„ ô
Wö9r4   rá   c                   óŠ   ‡ — e Zd ZdZˆ fd„Zdej                  dej                  dej                  dej                  fd„Zˆ xZS )ÚDots1Expertsz2Collection of expert weights stored as 3D tensors.c                 óò  •— t         ‰| �  «        |j                  | _        |j                  | _        |j                  | _        t        j                  t        j                  | j                  d| j                  z  | j
                  «      «      | _        t        j                  t        j                  | j                  | j
                  | j                  «      «      | _        t        |j                     | _        y )Nr7   )r)   r*   ræ   rç   r1   rè   Úmoe_intermediate_sizeÚintermediate_dimr   r+   r,   ÚemptyÚgate_up_projrÛ   r   rÜ   rÝ   rï   s     €r3   r*   zDots1Experts.__init__T  s­   ø€ Ü‰ÑÔØ!×3Ñ3ˆÔØ ×,Ñ,ˆŒØ &× <Ñ <ˆÔÜŸL™L¬¯©°T×5EÑ5EÀqÈ4×K`ÑK`ÑG`Ðbf×bqÑbqÓ)rÓsˆÔÜŸ™¤e§k¡k°$×2BÑ2BÀDÇOÁOÐUY×UjÑUjÓ&kÓlˆŒÜ˜V×.Ñ.Ñ/ˆ�r4   r5   Útop_k_indexÚtop_k_weightsr'   c                 óf  — t        j                  |«      }t        j                  «       5  t         j                  j                  j                  || j                  ¬«      }|j                  ddd«      }t        j                  |j                  d¬«      d«      j                  «       }d d d «       D ]ð  }|d   }|| j                  k(  rŒt        j                  |   «      \  }}	||	   }
t        j                  j                  |
| j                  |   «      j                  dd¬«      \  }}| j                  |«      |z  }t        j                  j                  || j                   |   «      }|||	|d f   z  }|j#                  d|	|j%                  |j&                  «      «       Œò |S # 1 sw Y   �ŒxY w)N)Únum_classesr7   r!   r   )r8   éþÿÿÿrr   r8   )r,   r÷   r…   r   r§   Úone_hotrç   ÚpermuteÚgreaterrö   ÚnonzeroÚwhereró   r  ÚchunkrÝ   rÛ   Ú
index_add_r;   r:   )r0   r5   r  r  Úfinal_hidden_statesÚexpert_maskÚ
expert_hitÚ
expert_idxÚ	top_k_posÚ	token_idxÚcurrent_stateÚgateÚupÚcurrent_hidden_statess                 r3   rB   zDots1Experts.forward]  s™  € ô $×.Ñ.¨}Ó=ÐÜ�]‰]‹_ñ 	SÜŸ(™(×-Ñ-×5Ñ5°kÈt×O_ÑO_Ð5Ó`ˆKØ%×-Ñ-¨a°°AÓ6ˆKÜŸ™ {§¡¸8 Ó'DÀaÓH×PÑPÓRˆJ÷	Sð
 %ò 
	nˆJØ# A™ˆJØ˜T×-Ñ-Ò-ØÜ#(§;¡;¨{¸:Ñ/FÓ#GÑ ˆI�yØ)¨)Ñ4ˆMÜ—}‘}×+Ñ+¨M¸4×;LÑ;LÈZÑ;XÓY×_Ñ_Ð`aÐgiÐ_Ój‰HˆD�"Ø$(§K¡K°Ó$5¸Ñ$:Ð!Ü$&§M¡M×$8Ñ$8Ð9NÐPT×P^ÑP^Ð_iÑPjÓ$kÐ!Ø$9¸MÈ)ÐU^Ð`dÐJdÑ<eÑ$eÐ!Ø×*Ñ*¨1¨iÐ9N×9QÑ9QÐRe×RkÑRkÓ9lÕmð
	nð #Ð"÷#	Sñ 	Sús   ªA=F&Æ&F0)	rG   rH   rI   rÓ   r*   r,   rK   rB   rL   rM   s   @r3   r  r  P  sF   ø„ á<ô0ð#à—|‘|ð#ð —\‘\ð#ð —|‘|ð	#ð
 
�‰÷#r4   r  c                   ó`   ‡ — 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Dots1MoEz:
    A mixed expert module containing shared experts.
    rQ   c                 óÂ   •— t         ‰| �  «        || _        t        |«      | _        t        |«      | _        t        ||j                  |j                  z  ¬«      | _
        y )N)rQ   rØ   )r)   r*   rQ   r  Úexpertsrá   r   rÕ   r	  Ún_shared_expertsÚshared_expertsrï   s     €r3   r*   zDots1MoE.__init__}  sP   ø€ Ü‰ÑÔØˆŒÜ# FÓ+ˆŒÜ# FÓ+ˆŒ	Ü&Ø¨V×-IÑ-IÈF×LcÑLcÑ-cô
ˆÕr4   r5   r'   c                 óò   — |}|j                   }| j                  |«      \  }}}|j                  d|j                   d   «      } | j                  |||«      j                  |Ž }|| j	                  |«      z   }|S )Nr8   )rE   r   rË   r&  r(  )r0   r5   Ú	residualsÚ
orig_shapeÚ_r  r  s          r3   rB   zDots1MoE.forward†  s   € Ø!ˆ	Ø"×(Ñ(ˆ
Ø(,¯	©	°-Ó(@Ñ%ˆˆ<˜Ø%×*Ñ*¨2¨}×/BÑ/BÀ2Ñ/FÓGˆØT˜Ÿ™ ]°LÀ,ÓO×TÑTÐV`ÐaˆØ%¨×(;Ñ(;¸IÓ(FÑFˆØÐr4   )
rG   rH   rI   rÓ   r"   r*   r,   rK   rB   rL   rM   s   @r3   r$  r$  x  s/   ø„ ñð
˜{õ 
ð U§\¡\ð °e·l±l÷ r4   r$  c                   ó  ‡ — e Zd Zdedefˆ fd„Z	 	 	 	 	 ddej                  dej                  dz  dej                  dz  de	dz  d	e
dz  d
eej                  ej                  f   dz  dee   dej                  fd„Zˆ xZS )ÚDots1DecoderLayerrQ   r±   c                 ó`  •— t         ‰| �  «        |j                  | _        t        ||¬«      | _        ||j
                  k\  rt        |«      | _        nt        |«      | _        t        |j                  |j                  ¬«      | _        t        |j                  |j                  ¬«      | _        y )N)rQ   r±   r¶   )r)   r*   r1   r°   Ú	self_attnÚfirst_k_dense_replacer$  ÚmlprÕ   r%   rÂ   Úinput_layernormÚpost_attention_layernormrÆ   s      €r3   r*   zDots1DecoderLayer.__init__‘  s†   ø€ Ü‰ÑÔØ!×-Ñ-ˆÔä'¨vÀÔKˆŒà˜×4Ñ4Ò4Ü Ó'ˆD�Hä Ó'ˆDŒHä+¨F×,>Ñ,>ÀF×DWÑDWÔXˆÔÜ(4°V×5GÑ5GÈV×M`ÑM`Ô(aˆÕ%r4   Nr5   rž   r|   rÈ   Ú	use_cacherÇ   r¡   r'   c           
      óÀ   — |}| j                  |«      } | j                  d||||||dœ|¤Ž\  }}	||z   }|}| j                  |«      }| j                  |«      }||z   }|S )N)r5   rž   r|   rÈ   r5  rÇ   © )r3  r0  r4  r2  )
r0   r5   rž   r|   rÈ   r5  rÇ   r¡   Úresidualr,  s
             r3   rB   zDots1DecoderLayer.forwardŸ  s�   € ð !ˆØ×,Ñ,¨]Ó;ˆà)˜4Ÿ>™>ð 
Ø'Ø)Ø%Ø+ØØ 3ñ
ð ñ
Ñˆ�qð ! =Ñ0ˆð !ˆØ×5Ñ5°mÓDˆØŸ™ Ó/ˆØ  =Ñ0ˆØÐr4   )NNNFN)rG   rH   rI   r"   r„   r*   r,   rK   Ú
LongTensorr	   rú   rD   r   r   rB   rL   rM   s   @r3   r.  r.  �  s¾   ø„ ðb˜{ð b°sõ bð" /3Ø04Ø(,Ø!&ØHLñà—|‘|ðð Ÿ™ tÑ+ðð ×&Ñ&¨Ñ-ð	ð
  ™ðð ˜$‘;ðð # 5§<¡<°·±Ð#=Ñ>ÀÑEðð Ð+Ñ,ðð 
�‰÷r4   r.  c                   óŽ   ‡ — e Zd ZU eed<   dZdZdgZdgZdZ	dZ
dZdZdZeedœZdgZdZ ej(                  «       ˆ fd	„«       Zˆ xZS )
ÚDots1PreTrainedModelrQ   ÚmodelTr.  rÈ   )r5   Ú
attentionsrã   Nc                 óê  •— t         ‰| �  |«       t        |t        «      rVt	        j
                  |j                  d| j                  j                  ¬«       t	        j                  |j                  «       y t        |t        «      rmt	        j
                  |j                  d| j                  j                  ¬«       t	        j
                  |j                  d| j                  j                  ¬«       y y )NrÊ   )r>   Ústd)r)   Ú_init_weightsrt   rá   ÚinitÚnormal_r.   rQ   Úinitializer_rangeÚzeros_rã   r  r  rÛ   )r0   rš   r2   s     €r3   r@  z"Dots1PreTrainedModel._init_weightsÓ  s™   ø€ ä‰Ñ˜fÔ%Ü�fœoÔ.Ü�L‰L˜Ÿ™¨S°d·k±k×6SÑ6SÕTÜ�K‰K˜×6Ñ6Õ7Ü˜¤Ô-Ü�L‰L˜×,Ñ,°3¸D¿K¹K×<YÑ<YÕZÜ�L‰L˜×)Ñ)°¸¿¹×9VÑ9VÖWð .r4   )rG   rH   rI   r"   r‚   Úbase_model_prefixÚsupports_gradient_checkpointingÚ_no_split_modulesÚ_skip_keys_device_placementÚ_supports_flash_attnÚ_supports_sdpaÚ_supports_flex_attnÚ_can_compile_fullgraphÚ_supports_attention_backendr.  r°   Ú_can_record_outputsÚ_keep_in_fp32_modules_strictÚ"_keys_to_ignore_on_load_unexpectedr,   r…   r@  rL   rM   s   @r3   r;  r;  ¿  s   ø… àÓØÐØ&*Ð#Ø,Ð-ÐØ#4Ð"5ÐØÐØ€NØÐà!ÐØ"&Ðà*Ø$ñÐð %>Ð#>Ð Ø)-Ð&à€U‡]�]ƒ_óXó ôXr4   r;  c                   óò   ‡ — e Zd Zdefˆ fd„Zeee	 	 	 	 	 	 ddej                  dz  dej                  dz  dej                  dz  dedz  dej                  dz  d	edz  d
ee   defd„«       «       «       Zˆ xZS )Ú
Dots1ModelrQ   c           	      óF  •— t         ‰| �  |«       |j                  | _        |j                  | _        t        j                  |j                  |j                  | j                  «      | _        t        j                  t        |j                  «      D �cg c]  }t        ||«      ‘Œ c}«      | _        t        |j                  |j                  ¬«      | _        t#        |¬«      | _        d| _        d| j(                  j*                  v | _        | j/                  «        y c c}w )Nr¶   ©rQ   Fr·   )r)   r*   Úpad_token_idÚpadding_idxÚ
vocab_sizer   Ú	Embeddingr1   Úembed_tokensÚ
ModuleListÚrangeÚnum_hidden_layersr.  Úlayersr%   rÂ   ÚnormrO   Ú
rotary_embÚgradient_checkpointingrQ   r³   Úhas_sliding_layersÚ	post_initrÆ   s      €r3   r*   zDots1Model.__init__à  sÝ   ø€ Ü‰Ñ˜Ô Ø!×.Ñ.ˆÔØ ×+Ñ+ˆŒäŸL™L¨×):Ñ):¸F×<NÑ<NÐPT×P`ÑP`ÓaˆÔÜ—m‘mÜCHÈ×IaÑIaÓCbÖc°iÔ˜v yÕ1Òcó
ˆŒô ! ×!3Ñ!3¸×9LÑ9LÔMˆŒ	Ü.°fÔ=ˆŒØ&+ˆÔ#Ø"5¸¿¹×9PÑ9PÐ"PˆÔð 	�‰Õùò ds   ÂDNÚ	input_idsrž   r|   rÈ   Úinputs_embedsr5  r¡   r'   c           
      óú  — |d u |d uz  rt        d«      ‚|€| j                  |«      }|r|€t        | j                  ¬«      }|€V|�|j	                  «       nd}t        j                  |j                  d   |j                  ¬«      |z   }|j                  d«      }t        |x}	t        «      s9| j                  ||||dœ}
dt        di |
¤Ži}	| j                  rt        di |
¤Ž|	d<   |}| j                  ||«      }t!        | j"                  d | j                  j$                   «      D ].  \  }} ||f|	| j                  j&                  |      ||||d	œ|¤Ž}Œ0 | j)                  |«      }t+        ||r|¬
«      S d ¬
«      S )Nz:You must specify exactly one of input_ids or inputs_embedsrT  r   r!   )r_   )rQ   rd  rž   rÈ   r|   Úfull_attentionr·   )rž   rÇ   r|   rÈ   r5  )Úlast_hidden_staterÈ   r7  )Ú
ValueErrorrY  r
   rQ   Úget_seq_lengthr,   rh   rE   r_   rŒ   rt   Údictr   ra  r   r_  Ú	enumerater]  r\  r³   r^  r   )r0   rc  rž   r|   rÈ   rd  r5  r¡   Úpast_seen_tokensÚcausal_mask_mappingÚmask_kwargsr5   rÇ   ÚiÚdecoder_layers                  r3   rB   zDots1Model.forwardñ  s¾  € ð ˜Ð -°tÐ";Ò<ÜÐYÓZÐZàÐ Ø ×-Ñ-¨iÓ8ˆMá˜Ð0Ü*°$·+±+Ô>ˆOàÐØCRÐC^˜×=Ñ=Ô?ÐdeÐÜ Ÿ<™<¨×(;Ñ(;¸AÑ(>À}×G[ÑG[Ô\Ð_oÑoˆLØ'×1Ñ1°!Ó4ˆLô °Ð?Ð-ÄÔFð Ÿ+™+Ø!.Ø"0Ø#2Ø ,ñˆKð !Ô"4Ñ"C°{Ñ"Cð#Ðð ×&Ò&Ü;\Ñ;kÐ_jÑ;kÐ#Ð$7Ñ8à%ˆØ"Ÿo™o¨m¸\ÓJÐä )¨$¯+©+Ð6U¸¿¹×8UÑ8UÐ*VÓ Wò 		ÑˆAˆ}Ù)Øðà2°4·;±;×3JÑ3JÈ1Ñ3MÑNØ$7Ø)Ø /Ø#ñð ñ‰Mð		ð Ÿ	™	 -Ó0ˆÜ&Ø+Ù/8˜Oô
ð 	
à>Bô
ð 	
r4   )NNNNNN)rG   rH   rI   r"   r*   r   r    r   r,   r9  rK   r	   ÚFloatTensorrú   r   r   r   rB   rL   rM   s   @r3   rR  rR  Þ  sÍ   ø„ ð˜{õ ð"  ØØð .2Ø.2Ø04Ø(,Ø26Ø!%ñ<
à×#Ñ# dÑ*ð<
ð Ÿ™ tÑ+ð<
ð ×&Ñ&¨Ñ-ð	<
ð
  ™ð<
ð ×(Ñ(¨4Ñ/ð<
ð ˜$‘;ð<
ð Ð+Ñ,ð<
ð 
!ò<
ó ó ó  ô<
r4   rR  c                   óB  ‡ — e Zd ZddiZddiZddgdgfiZˆ fd„Zee	 	 	 	 	 	 	 	 dd	e	j                  dz  d
e	j                  dz  de	j                  dz  dedz  de	j                  dz  de	j                  dz  dedz  dee	j                  z  dee   defd„«       «       Zˆ xZS )ÚDots1ForCausalLMzlm_head.weightzmodel.embed_tokens.weightÚlm_headÚcolwise_gather_outputr5   Úlogitsc                 óè   •— t         ‰| �  |«       t        |«      | _        |j                  | _        t        j                  |j                  |j                  d¬«      | _        | j                  «        y r×   )
r)   r*   rR  r<  rW  r   r¼   r1   rt  rb  rï   s     €r3   r*   zDots1ForCausalLM.__init__9  sU   ø€ Ü‰Ñ˜Ô Ü Ó'ˆŒ
Ø ×+Ñ+ˆŒÜ—y‘y ×!3Ñ!3°V×5FÑ5FÈUÔSˆŒð 	�‰Õr4   Nrc  rž   r|   rÈ   rd  Úlabelsr5  Úlogits_to_keepr¡   r'   c	           
      óx  —  | j                   d||||||dœ|	¤Ž}
|
j                  }t        |t        «      rt	        | d«      n|}| j                  |dd…|dd…f   «      }d}|�* | j                  d||| j                  j                  dœ|	¤Ž}t        |||
j                  |
j                  |
j                  ¬«      S )a~  
        labels (`torch.LongTensor` of shape `(batch_size, sequence_length)`, *optional*):
            Labels for computing the masked language modeling loss. Indices should either be in `[0, ...,
            config.vocab_size]` or -100 (see `input_ids` docstring). Tokens with indices set to `-100` are ignored
            (masked), the loss is only computed for the tokens with labels in `[0, ..., config.vocab_size]`.

        Example:

        ```python
        >>> from transformers import AutoTokenizer, Dots1ForCausalLM

        >>> model = Dots1ForCausalLM.from_pretrained("rednote-hilab/dots1.llm1.inst")
        >>> tokenizer = AutoTokenizer.from_pretrained("rednote-hilab/dots1.llm1.inst")

        >>> prompt = "Hey, are you conscious? Can you talk to me?"
        >>> inputs = tokenizer(prompt, return_tensors="pt")

        >>> # Generate
        >>> generate_ids = model.generate(inputs.input_ids, max_length=30)
        >>> tokenizer.batch_decode(generate_ids, skip_special_tokens=True, clean_up_tokenization_spaces=False)[0]
        "Hey, are you conscious? Can you talk to me?\nI'm not conscious, but I can talk to you."
        ```)rc  rž   r|   rÈ   rd  r5  N)rv  rx  rW  )Úlossrv  rÈ   r5   r=  r7  )r<  rg  rt   r„   Úslicert  Úloss_functionrQ   rW  r   rÈ   r5   r=  )r0   rc  rž   r|   rÈ   rd  rx  r5  ry  r¡   Úoutputsr5   Úslice_indicesrv  r{  s                  r3   rB   zDots1ForCausalLM.forwardB  sÚ   € ðH ,6¨4¯:©:ð ,
ØØ)Ø%Ø+Ø'Øñ,
ð ñ,
ˆð  ×1Ñ1ˆä8BÀ>ÔSVÔ8Wœ˜~˜o¨tÔ4Ð]kˆØ—‘˜mªA¨}ºaÐ,?Ñ@ÓAˆàˆØÐØ%�4×%Ñ%Ðp¨V¸FÈtÏ{É{×OeÑOeÑpÐioÑpˆDä%ØØØ#×3Ñ3Ø!×/Ñ/Ø×)Ñ)ô
ð 	
r4   )NNNNNNNr   )rG   rH   rI   Ú_tied_weights_keysÚ_tp_planÚ_pp_planr*   r   r   r,   r9  rK   r	   rq  rú   r„   r   r   r   rB   rL   rM   s   @r3   rs  rs  3  s  ø„ à*Ð,GÐHÐØÐ2Ð3€HØ˜_Ð-°¨zÐ:Ð;€Hôð Øð .2Ø.2Ø04Ø(,Ø26Ø*.Ø!%Ø-.ñ;
à×#Ñ# dÑ*ð;
ð Ÿ™ tÑ+ð;
ð ×&Ñ&¨Ñ-ð	;
ð
  ™ð;
ð ×(Ñ(¨4Ñ/ð;
ð × Ñ  4Ñ'ð;
ð ˜$‘;ð;
ð ˜eŸl™lÑ*ð;
ð Ð+Ñ,ð;
ð 
 ò;
ó ó ô;
r4   rs  )r;  rR  rs  )r!   )rÊ   )IÚcollections.abcr   Útypingr   r,   Útorch.nn.functionalr   r§   rò   Ú r   rA  Úactivationsr   Úcache_utilsr	   r
   Ú
generationr   Úintegrationsr   r   r   r   Úmasking_utilsr   r   Úmodeling_flash_attention_utilsr   Úmodeling_layersr   Úmodeling_outputsr   r   Úmodeling_rope_utilsr   r   Úmodeling_utilsr   r   Úprocessing_utilsr   Úutilsr   r   r   Úutils.genericr   r   Úutils.output_capturingr    Úconfiguration_dots1r"   ÚModuler%   rO   r‰   r’   rK   r„   r™   rJ   r®   r°   rÕ   rá   r  r$  r.  r;  rR  rs  Ú__all__r7  r4   r3   ú<module>r˜     s8  ðõ( %Ý ã ß Ð Ý å &Ý !ß .Ý )÷ó ÷ SÝ BÝ 9ß Oß Kß FÝ &ß IÑ Iß GÝ 5Ý ,ñ ˜YÓ'ôJ�2—9‘9ó Jó (ðJô(><˜2Ÿ9™9ô ><òB(ñ Ð*Ó+òó ,ðð2	U˜UŸ\™\ð 	U°#ð 	U¸%¿,¹,ó 	Uð& ñ%Ø�I‰Ið%à�<‰<ð%ð 
�‰ð%ð �<‰<ð	%ð
 —L‘L 4Ñ'ð%ð ð%ð ð%ð Ð'Ñ(ó%ñ2 Ð)Ó*ôE)�R—Y‘Yó E)ó +ðE)ôPˆr�y‰yô ô &9�b—i‘iô &9ðR ô$#�2—9‘9ó $#ó ð$#ôNˆr�y‰yô ô0,Ð2ô ,ð^ ôX˜?ó Xó ðXð< ôQ
Ð%ó Q
ó ðQ
ðh ôK
Ð+¨_ó K
ó ðK
ò\ E�r4   