Ë
    ÛÓ^j²X  ã                   ó\  — d dl mZ d dlmZ d dl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 dd	lmZ dd
lmZ ddlmZmZ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(m)Z) ddl*m+Z+ ddl,m-Z-  G d„ dej\                  «      Z/ G d„ de«      Z0dejb                  de2dejb                  fd„Z3	 d:dej\                  dejb                  dejb                  d ejb                  d!ejb                  dz  d"e4d#e4d$e"e$   fd%„Z5d&„ Z6d;d'„Z7 G d(„ d)ej\                  «      Z8 G d*„ d+ej\                  «      Z9 ed,«       G d-„ d.ej\                  «      «       Z:e% G d/„ d0e «      «       Z;e% G d1„ d2e;«      «       Z<e% G d3„ d4e;e«      «       Z= G d5„ d6ee;«      Z> G d7„ d8ee;«      Z?g d9¢Z@y)<é    )ÚCallable)ÚOptionalNé   )ÚACT2FN)ÚCacheÚDynamicCache)ÚGenerationMixin)Úuse_kernel_forward_from_hub)Úcreate_causal_mask)ÚFlashAttentionKwargs)Ú GenericForSequenceClassificationÚGenericForTokenClassificationÚ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
Glm4Configc                   óV   ‡ — e Zd Zˆ fd„Zdej
                  dej
                  fd„Zˆ xZS )ÚGlm4MLPc                 ó*  •— t         ‰| �  «        || _        t        j                  |j
                  d|j                  z  d¬«      | _        t        j                  |j                  |j
                  d¬«      | _        t        |j                     | _        y )Né   F©Úbias)ÚsuperÚ__init__ÚconfigÚnnÚLinearÚhidden_sizeÚintermediate_sizeÚgate_up_projÚ	down_projr   Ú
hidden_actÚactivation_fn©Úselfr'   Ú	__class__s     €úq/var/www/ramen.bs-engineer-server.com/venv/lib/python3.12/site-packages/transformers/models/glm4/modeling_glm4.pyr&   zGlm4MLP.__init__2   sp   ø€ Ü‰ÑÔàˆŒÜŸI™I f×&8Ñ&8¸!¸f×>VÑ>VÑ:VÐ]bÔcˆÔÜŸ™ 6×#;Ñ#;¸V×=OÑ=OÐV[Ô\ˆŒÜ# F×$5Ñ$5Ñ6ˆÕó    Úhidden_statesÚreturnc                 óš   — | j                  |«      }|j                  dd¬«      \  }}|| j                  |«      z  }| j                  |«      S )Nr"   éÿÿÿÿ©Údim)r,   Úchunkr/   r-   )r1   r5   Ú	up_statesÚgates       r3   ÚforwardzGlm4MLP.forward:   sL   € Ø×%Ñ% mÓ4ˆ	à#Ÿ/™/¨!°˜/Ó4‰ˆˆiØ × 2Ñ 2°4Ó 8Ñ8ˆ	à�~‰~˜iÓ(Ð(r4   )Ú__name__Ú
__module__Ú__qualname__r&   ÚtorchÚFloatTensorr>   Ú__classcell__©r2   s   @r3   r    r    1   s'   ø„ ô7ð) U×%6Ñ%6ð )¸5×;LÑ;L÷ )r4   r    c                   óD  ‡ — 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ej                  eej                  ej                  f   dz  f   fd„Zˆ xZS )ÚGlm4DecoderLayerr'   Ú	layer_idxc                 ó¸  •— t         ‰| �  «        |j                  | _        t        ||¬«      | _        t        |«      | _        t        |j                  |j                  ¬«      | _	        t        |j                  |j                  ¬«      | _
        t        |j                  |j                  ¬«      | _        t        |j                  |j                  ¬«      | _        y )N)r'   rH   ©Úeps)r%   r&   r*   ÚGlm4AttentionÚ	self_attnr    ÚmlpÚGlm4RMSNormÚrms_norm_epsÚinput_layernormÚpost_attention_layernormÚpost_self_attn_layernormÚpost_mlp_layernorm©r1   r'   rH   r2   s      €r3   r&   zGlm4DecoderLayer.__init__D   s¡   ø€ Ü‰ÑÔØ!×-Ñ-ˆÔÜ&¨fÀ	ÔJˆŒä˜6“?ˆŒÜ*¨6×+=Ñ+=À6×CVÑCVÔWˆÔÜ(3°F×4FÑ4FÈF×L_ÑL_Ô(`ˆÔ%Ü(3°F×4FÑ4FÈF×L_ÑL_Ô(`ˆÔ%Ü"-¨f×.@Ñ.@Àf×FYÑFYÔ"ZˆÕr4   Nr5   Úattention_maskÚposition_idsÚpast_key_valuesÚ	use_cacheÚposition_embeddingsÚkwargsr6   c           
      ó  — |}| j                  |«      } | j                  d||||||dœ|¤Ž\  }}	| j                  |«      }||z   }|}| j                  |«      }| j	                  |«      }| j                  |«      }||z   }|S )N)r5   rV   rW   rX   rY   rZ   © )rQ   rM   rS   rR   rN   rT   )
r1   r5   rV   rW   rX   rY   rZ   r[   ÚresidualÚ_s
             r3   r>   zGlm4DecoderLayer.forwardO   s®   € ð !ˆØ×,Ñ,¨]Ó;ˆà)˜4Ÿ>™>ð 
Ø'Ø)Ø%Ø+ØØ 3ñ
ð ñ
Ñˆ�qð ×5Ñ5°mÓDˆØ  =Ñ0ˆà ˆØ×5Ñ5°mÓDˆØŸ™ Ó/ˆØ×/Ñ/°Ó>ˆØ  =Ñ0ˆØÐr4   )NNNFN)r?   r@   rA   r   Úintr&   rB   ÚTensorÚ
LongTensorr   ÚboolÚtupler   r   rC   r>   rD   rE   s   @r3   rG   rG   C   sæ   ø„ ð	[˜zð 	[°cõ 	[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 
ˆu× Ñ  %¨×(9Ñ(9¸5×;LÑ;LÐ(LÑ"MÐPTÑ"TÐTÑ	U÷r4   rG   r5   Ún_repr6   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)ÚshapeÚexpandÚreshape)r5   re   ÚbatchÚnum_key_value_headsÚslenÚhead_dims         r3   Ú	repeat_kvrn   q   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valuerV   ÚscalingÚdropoutr[   c                 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 )Nr"   r   r8   )r:   Údtype)ÚpÚtrainingr   )rn   Únum_key_value_groupsrB   ÚmatmulÚ	transposer(   Ú
functionalÚsoftmaxÚfloat32Útorv   rt   rx   Ú
contiguous)ro   rp   rq   rr   rV   rs   rt   r[   Ú
key_statesÚvalue_statesÚattn_weightsÚattn_outputs               r3   Úeager_attention_forwardr…   }   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                 ó|   — | dddd…f   }| dddd…f   }t        j                  | |fd¬«      j                  d«      S )	z*Rotates half the hidden dims of the input..r   Nr"   r   r8   r9   éþÿÿÿ)rB   ÚstackÚflatten)ÚxÚx1Úx2s      r3   Úrotate_halfr�   –   sJ   € à	
ˆ3���1�ˆ9‰€BØ	
ˆ3���1�ˆ9‰€BÜ�;‰;˜˜˜R�y bÔ)×1Ñ1°"Ó5Ð5r4   c                 ó  — |j                  |«      }|j                  |«      }|dd|j                  d   dz  …f   j                  dd¬«      }|dd|j                  d   dz  …f   j                  dd¬«      }|j                  d   }| dd|…f   | d|d…f   }}|dd|…f   |d|d…f   }	}||z  t        |«      |z  z   }
||z  t        |«      |z  z   }t	        j
                  |
|gd¬«      }
t	        j
                  ||	gd¬«      }|
|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.
    .Nr8   r"   r9   )Ú	unsqueezerg   Úrepeat_interleaver�   rB   Úcat)ÚqÚkÚcosÚsinÚunsqueeze_dimÚ
rotary_dimÚq_rotÚq_passÚk_rotÚk_passÚq_embedÚk_embeds               r3   Úapply_rotary_pos_embrž   �   sD  € ð$ �-‰-˜Ó
&€CØ
�-‰-˜Ó
&€Cð ˆcÐ'�S—Y‘Y˜r‘] aÑ'Ð'Ð'Ñ
(×
:Ñ
:¸1À"Ð
:Ó
E€CØ
ˆcÐ'�S—Y‘Y˜r‘] aÑ'Ð'Ð'Ñ
(×
:Ñ
:¸1À"Ð
:Ó
E€Cð —‘˜2‘€JØ�c˜;˜J˜;Ð&Ñ'¨¨3°
±Ð+;Ñ)<ˆ6€EØ�c˜;˜J˜;Ð&Ñ'¨¨3°
±Ð+;Ñ)<ˆ6€Eð �s‰{œ{¨5Ó1°CÑ7Ñ8€GØ�s‰{œ{¨5Ó1°CÑ7Ñ8€Gô �i‰i˜ &Ð)¨rÔ2€GÜ�i‰i˜ &Ð)¨rÔ2€GØ�GÐÐr4   c                   ó   ‡ — e Zd ZdZddededz  fˆ fd„Z	 	 	 ddej                  de	ej                  ej                  f   dz  dej                  dz  d	e
dz  d
ee   de	ej                  ej                  f   fd„Zˆ xZS )rL   z=Multi-headed attention from 'Attention Is All You Need' paperNr'   rH   c                 óP  •— t         ‰| �  «        || _        || _        t	        |d|j
                  |j                  z  «      | _        |j                  |j                  z  | _	        | j                  dz  | _
        |j                  | _        d| _        t        j                  |j
                  |j                  | j                  z  |j                  ¬«      | _        t        j                  |j
                  |j                  | j                  z  |j                  ¬«      | _        t        j                  |j
                  |j                  | j                  z  |j                  ¬«      | _        t        j                  |j                  | j                  z  |j
                  d¬«      | _        y )Nrm   g      à¿Tr#   F)r%   r&   r'   rH   Úgetattrr*   Únum_attention_headsrm   rk   ry   rs   Úattention_dropoutÚ	is_causalr(   r)   Úattention_biasÚq_projÚk_projÚv_projÚo_projrU   s      €r3   r&   zGlm4Attention.__init__È   sD  ø€ Ü‰ÑÔØˆŒØ"ˆŒÜ ¨
°F×4FÑ4FÈ&×JdÑJdÑ4dÓeˆŒØ$*×$>Ñ$>À&×B\ÑB\Ñ$\ˆÔ!Ø—}‘} dÑ*ˆŒØ!'×!9Ñ!9ˆÔØˆŒä—i‘iØ×Ñ × :Ñ :¸T¿]¹]Ñ JÐQW×QfÑQfô
ˆŒô —i‘iØ×Ñ × :Ñ :¸T¿]¹]Ñ JÐQW×QfÑQfô
ˆŒô —i‘iØ×Ñ × :Ñ :¸T¿]¹]Ñ JÐQW×QfÑQfô
ˆŒô —i‘i × :Ñ :¸T¿]¹]Ñ JÈF×L^ÑL^ÐejÔkˆ�r4   r5   rZ   rV   rX   r[   r6   c                 ó
  — |j                   d d }g |¢d‘| j                  ‘­}| j                  |«      j                  |«      j	                  dd«      }| j                  |«      j                  |«      j	                  dd«      }	| j                  |«      j                  |«      j	                  dd«      }
|\  }}t        ||	||«      \  }}	|� |j                  |	|
| j                  «      \  }	}
t        j                  | j                  j                  t        «      } || ||	|
|f| j                  sdn| j                   | j"                  dœ|¤Ž\  }} |j$                  g |¢d‘­Ž j'                  «       }| j)                  |«      }||fS )Nr8   r   r"   ç        )rt   rs   )rg   rm   r¦   Úviewr{   r§   r¨   rž   ÚupdaterH   r   Úget_interfacer'   Ú_attn_implementationr…   rx   r£   rs   ri   r€   r©   )r1   r5   rZ   rV   rX   r[   Úinput_shapeÚhidden_shapeÚquery_statesr�   r‚   r”   r•   Úattention_interfacer„   rƒ   s                   r3   r>   zGlm4Attention.forwardÝ   s‘  € ð $×)Ñ)¨#¨2Ð.ˆØ8˜Ð8 bÐ8¨$¯-©-Ñ8ˆà—{‘{ =Ó1×6Ñ6°|ÓD×NÑNÈqÐRSÓTˆØ—[‘[ Ó/×4Ñ4°\ÓB×LÑLÈQÐPQÓRˆ
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‘Lñ	%
ð ñ	%
Ñ!ˆ�\ð *�k×)Ñ)Ð;¨;Ð;¸Ò;×FÑFÓHˆØ—k‘k +Ó.ˆØ˜LÐ(Ð(r4   ©N©NNN)r?   r@   rA   Ú__doc__r   r`   r&   rB   ra   rd   r   r   r   r>   rD   rE   s   @r3   rL   rL   Å   s¬   ø„ ÙGñl˜zð l°c¸D±jõ lð0 IMØ.2Ø(,ñ&)à—|‘|ð&)ð # 5§<¡<°·±Ð#=Ñ>ÀÑEð&)ð Ÿ™ tÑ+ð	&)ð
  ™ð&)ð Ð+Ñ,ð&)ð 
ˆu�|‰|˜UŸ\™\Ð)Ñ	*÷&)r4   rL   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Glm4RotaryEmbeddingÚinv_freqNr'   c                 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defaultr¹   F)Ú
persistentÚoriginal_inv_freq)r%   r&   Úmax_position_embeddingsÚmax_seq_len_cachedÚoriginal_max_seq_lenr'   Úrope_parametersr»   Úcompute_default_rope_parametersr   Úattention_scalingÚregister_bufferÚclone)r1   r'   ÚdeviceÚrope_init_fnr¹   r2   s        €r3   r&   zGlm4RotaryEmbedding.__init__	  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   rÇ   ztorch.deviceÚseq_lenr6   ztorch.Tensorc                 ón  — | j                   d   }| j                   j                  dd«      }t        | dd«      xs | j                  | j                  z  }t        ||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partial_rotary_factorg      ð?rm   Nr   r"   ©rv   )rÇ   rv   )rÂ   Úgetr¡   r*   r¢   r`   rB   ÚarangeÚint64r   Úfloat)	r'   rÇ   rÉ   ÚbaserÌ   rm   r:   Úattention_factorr¹   s	            r3   rÃ   z3Glm4RotaryEmbedding.compute_default_rope_parameters  s»   € ð& ×%Ñ% lÑ3ˆØ &× 6Ñ 6× :Ñ :Ð;RÐTWÓ XÐÜ˜6 :¨tÓ4Òh¸×8JÑ8JÈf×NhÑNhÑ8hˆÜ�(Ð2Ñ2Ó3ˆ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enabledr"   r9   rÍ   )r¹   rÑ   rh   rg   r   rÇ   Ú
isinstanceÚtypeÚstrr   r{   rB   r‘   r”   rÄ   r•   rv   )
r1   rŠ   rW   Úinv_freq_expandedÚposition_ids_expandedr×   ÚfreqsÚembr”   r•   s
             r3   r>   zGlm4RotaryEmbedding.forward9  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$r´   rµ   )r?   r@   rA   rB   ra   Ú__annotations__r   r&   Ústaticmethodr   r`   rd   rÑ   rÃ   Úno_gradr   r>   rD   rE   s   @r3   r¸   r¸     s˜   ø… Ø�l‰lÓñV˜zõ Vð  à$(Ø+/Ø"ñ*Ø˜TÑ!ð*à˜Ñ(ð*ð �t‘ð*ð 
ˆ~˜uÐ$Ñ	%ò	*ó ð*ð> €U‡]�]ƒ_Øñ<ó ó ô<r4   r¸   ÚRMSNormc                   óh   ‡ — e Zd Zddeddfˆ fd„Zdej                  dej                  fd„Zd„ Zˆ xZ	S )	rO   rK   r6   Nc                 óŠ   •— t         ‰| �  «        t        j                  t	        j
                  |«      «      | _        || _        y)z:
        Glm4RMSNorm is equivalent to T5LayerNorm
        N)r%   r&   r(   Ú	ParameterrB   ÚonesÚweightÚvariance_epsilon)r1   r*   rK   r2   s      €r3   r&   zGlm4RMSNorm.__init__K  s1   ø€ ô 	‰ÑÔÜ—l‘l¤5§:¡:¨kÓ#:Ó;ˆŒØ #ˆÕr4   r5   c                 ó"  — |j                   }|j                  t        j                  «      }|j	                  d«      j                  dd¬«      }|t        j                  || j                  z   «      z  }| j                  |j                  |«      z  S )Nr"   r8   T)Úkeepdim)	rv   r   rB   r~   ÚpowÚmeanÚrsqrtré   rè   )r1   r5   Úinput_dtypeÚvariances       r3   r>   zGlm4RMSNorm.forwardS  sy   € Ø#×)Ñ)ˆØ%×(Ñ(¬¯©Ó7ˆØ ×$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=)rd   rè   rg   ré   )r1   s    r3   Ú
extra_reprzGlm4RMSNorm.extra_reprZ  s*   € Ü˜Ÿ™×)Ñ)Ó*Ð+¨6°$×2GÑ2GÐ1HÐIÐIr4   )g�íµ ÷Æ°>)
r?   r@   rA   rÑ   r&   rB   ra   r>   rò   rD   rE   s   @r3   rO   rO   I  s7   ø„ ñ$¨ð $¸$õ $ð; U§\¡\ð ;°e·l±ló ;öJr4   rO   c                   óJ   — 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y)ÚGlm4PreTrainedModelr'   ÚmodelTrG   rX   )r5   Ú
attentionsN)r?   r@   rA   r   rà   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_backendrG   rL   Ú_can_record_outputsr]   r4   r3   rô   rô   ^  sQ   … àÓØÐØ&*Ð#Ø+Ð,ÐØ#4Ð"5ÐØÐØ€NØÐà!ÐØ"&Ðà)Ø#ñÑr4   rô   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	Glm4Modelr'   c           	      ó  •— t         ‰| �  |«       |j                  | _        |j                  | _        t        j                  |j                  |j                  | j                  «      | _        t        j                  t        |j                  «      D �cg c]  }t        ||«      ‘Œ c}«      | _        t        |j                  |j                  ¬«      | _        t#        |¬«      | _        d| _        | j)                  «        y c c}w )NrJ   ©r'   F)r%   r&   Úpad_token_idÚpadding_idxÚ
vocab_sizer(   Ú	Embeddingr*   Úembed_tokensÚ
ModuleListÚrangeÚnum_hidden_layersrG   ÚlayersrO   rP   Únormr¸   Ú
rotary_embÚgradient_checkpointingÚ	post_initrU   s      €r3   r&   zGlm4Model.__init__s  sÆ   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‘mÜBGÈ×H`ÑH`ÓBaÖb°YÔ˜f iÕ0Òbó
ˆŒô   × 2Ñ 2¸×8KÑ8KÔLˆŒ	Ü-°VÔ<ˆŒØ&+ˆÔ#ð 	�‰Õùò cs   ÂDNÚ	input_idsrV   rW   rX   Úinputs_embedsrY   r[   r6   c           
      ó<  — |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        | j                  ||||¬«      }	|}
| j                  |
|¬«      }| j                  d | j                  j                   D ]  } ||
f|	||||dœ|¤Ž}
Œ | j                  |
«      }
t        |
|¬	«      S )
Nz:You must specify exactly one of input_ids or inputs_embedsr  r   r   )rÇ   )r'   r  rV   rX   rW   )rW   )rV   rZ   rW   rX   rY   )Úlast_hidden_staterX   )Ú
ValueErrorr	  r   r'   Úget_seq_lengthrB   rÏ   rg   rÇ   r�   r   r  r  r  r  r   )r1   r  rV   rW   rX   r  rY   r[   Úpast_seen_tokensÚcausal_maskr5   rZ   Údecoder_layers                r3   r>   zGlm4Model.forwardƒ  sL  € ð ˜Ð -°tÐ";Ò<ÜÐYÓZÐZàÐ Ø*.×*;Ñ*;¸IÓ*FˆMá˜Ð0Ü*°$·+±+Ô>ˆOàÐØCRÐC^˜×=Ñ=Ô?ÐdeÐÜ Ÿ<™<¨×(;Ñ(;¸AÑ(>À}×G[ÑG[Ô\Ð_oÑoˆLØ'×1Ñ1°!Ó4ˆLä(Ø—;‘;Ø'Ø)Ø+Ø%ô
ˆð &ˆØ"Ÿo™o¨mÈ,˜oÓWÐà!Ÿ[™[Ð)H¨4¯;©;×+HÑ+HÐIò 		ˆMÙ)Øðà*Ø$7Ø)Ø /Ø#ñð ñ‰Mð		ð Ÿ	™	 -Ó0ˆÜ&Ø+Ø+ô
ð 	
r4   )NNNNNN)r?   r@   rA   r   r&   r   r   r   rB   rb   ra   r   rC   rc   r   r   r   r>   rD   rE   s   @r3   r  r  q  sÍ   ø„ ð˜zõ ð   ØØð .2Ø.2Ø04Ø(,Ø26Ø!%ñ2
à×#Ñ# dÑ*ð2
ð Ÿ™ tÑ+ð2
ð ×&Ñ&¨Ñ-ð	2
ð
  ™ð2
ð ×(Ñ(¨4Ñ/ð2
ð ˜$‘;ð2
ð Ð+Ñ,ð2
ð 
!ò2
ó ó ó  ô2
r4   r  c                   óH  ‡ — 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ez  fd„«       «       Zˆ xZS )ÚGlm4ForCausalLMz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 )NFr#   )
r%   r&   r  rõ   r  r(   r)   r*   r  r  r0   s     €r3   r&   zGlm4ForCausalLM.__init__Á  sU   ø€ Ü‰Ñ˜Ô Ü˜vÓ&ˆŒ
Ø ×+Ñ+ˆŒÜ—y‘y ×!3Ñ!3°V×5FÑ5FÈUÔSˆŒð 	�‰Õr4   Nr  rV   rW   rX   r  ÚlabelsrY   Úlogits_to_keepr[   r6   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 )ah  
        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, Glm4ForCausalLM

        >>> model = Glm4ForCausalLM.from_pretrained("THUDM/GLM-4-9B-0414")
        >>> tokenizer = AutoTokenizer.from_pretrained("THUDM/GLM-4-9B-0414")

        >>> 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."
        ```)r  rV   rW   rX   r  rY   N)r  r!  r  )Úlossr  rX   r5   rö   r]   )rõ   r  rÙ   r`   Úslicer  Úloss_functionr'   r  r   rX   r5   rö   )r1   r  rV   rW   rX   r  r!  rY   r"  r[   Úoutputsr5   Úslice_indicesr  r$  s                  r3   r>   zGlm4ForCausalLM.forwardÊ  sÚ   € ðH ,6¨4¯:©:ð ,
ØØ)Ø%Ø+Ø'Øñ,
ð ñ,
ˆð  ×1Ñ1ˆä8BÀ>ÔSVÔ8Wœ˜~˜o¨tÔ4Ð]kˆØ—‘˜mªA¨}ºaÐ,?Ñ@ÓAˆàˆØÐØ%�4×%Ñ%Ðp¨V¸FÈtÏ{É{×OeÑOeÑpÐioÑpˆDä%ØØØ#×3Ñ3Ø!×/Ñ/Ø×)Ñ)ô
ð 	
r4   )NNNNNNNr   )r?   r@   rA   Ú_tied_weights_keysÚ_tp_planÚ_pp_planr&   r   r   rB   rb   ra   r   rC   rc   r`   r   r   rd   r   r>   rD   rE   s   @r3   r  r  »  s  ø„ à*Ð,GÐHÐØÐ2Ð3€HØ˜_Ð-°¨zÐ:Ð;€Hôð Øð .2Ø.2Ø04Ø(,Ø26Ø*.Ø!%Ø-.ñ;
à×#Ñ# dÑ*ð;
ð Ÿ™ tÑ+ð;
ð ×&Ñ&¨Ñ-ð	;
ð
  ™ð;
ð ×(Ñ(¨4Ñ/ð;
ð × Ñ  4Ñ'ð;
ð ˜$‘;ð;
ð ˜eŸl™lÑ*ð;
ð Ð+Ñ,ð;
ð 
Ð'Ñ	'ò;
ó ó ô;
r4   r  c                   ó   — e Zd Zy)ÚGlm4ForSequenceClassificationN©r?   r@   rA   r]   r4   r3   r-  r-  
  ó   „ Ør4   r-  c                   ó   — e Zd Zy)ÚGlm4ForTokenClassificationNr.  r]   r4   r3   r1  r1    r/  r4   r1  )rô   r  r  r-  r1  )r«   )r   )AÚcollections.abcr   Útypingr   rB   Útorch.nnr(   Úactivationsr   Úcache_utilsr   r   Ú
generationr	   Úintegrationsr
   Úmasking_utilsr   Úmodeling_flash_attention_utilsr   Úmodeling_layersr   r   r   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_glm4r   ÚModuler    rG   ra   r`   rn   rÑ   r…   r�   rž   rL   r¸   rO   rô   r  r  r-  r1  Ú__all__r]   r4   r3   ú<module>rF     sà  ðõ, %Ý ã Ý å !ß .Ý )Ý 7Ý /Ý B÷ñ ÷
 Pß Kß FÝ &ß IÑ Iß GÝ 5Ý *ô)ˆb�i‰iô )ô$+Ð1ô +ð\	U˜UŸ\™\ð 	U°#ð 	U¸%¿,¹,ó 	Uð& ñ%Ø�I‰Ið%à�<‰<ð%ð 
�‰ð%ð �<‰<ð	%ð
 —L‘L 4Ñ'ð%ð ð%ð ð%ð Ð'Ñ(ó%ò26ó%ôP>)�B—I‘Iô >)ôB@<˜"Ÿ)™)ô @<ñF ˜YÓ'ôJ�"—)‘)ó Jó (ðJð( ô˜/ó ó ðð$ ôF
Ð#ó F
ó ðF
ðR ôK
Ð)¨?ó K
ó ðK
ô\	Ð$DÐFYô 	ô	Ð!>Ð@Sô 	ò�r4   