
    鬜iv                        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
 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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`                        Z1 ed       G d dej`                               Z2 G d dej`                        Z3 G d dej`                        Z4 G d d ejj                        Z6 G d! d"ej`                        Z7d# Z8 ed$      d>d%       Z9d&ejt                  d'e;d(ejt                  fd)Z<	 d?d*ej`                  d+ejt                  d,ejt                  d-ejt                  d.ejt                  dz  d/e=d0e=d1e$e&   fd2Z> ee9       G d3 d4ej`                               Z? G d5 d6e      Z@ G d7 d8e"      ZAe' G d9 d:eA             ZBe' G d; d<eAe             ZCg d=ZDy)@    )Callable)OptionalN)nn   )initialization)ACT2FN)CacheDynamicCache)GenerationMixin)use_kernel_forward_from_hubuse_kernel_func_from_hubuse_kernelized_func)create_causal_mask!create_sliding_window_causal_mask)GradientCheckpointingLayer)BaseModelOutputWithPastCausalLMOutputWithPastMoeModelOutputWithPast)ROPE_INIT_FUNCTIONSdynamic_rope_update)ALL_ATTENTION_FUNCTIONSPreTrainedModel)Unpack)TransformersKwargsauto_docstringcan_return_tuple)maybe_autocastmerge_with_config_defaults)capture_outputs   )AfmoeConfigc                        e Zd ZU ej                  ed<   ddef 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 xZS )AfmoeRotaryEmbeddinginv_freqNconfigc                    t         |           |j                  | _        |j                  | _        || _        | j
                  j                  d   | _        | j                  }| j                  dk7  rt        | j                     } || j
                  |      \  }| _
        | j                  d|d       | j                  d|j                         d       y )N	rope_typedefaultr$   F)
persistentoriginal_inv_freq)super__init__max_position_embeddingsmax_seq_len_cachedoriginal_max_seq_lenr%   rope_parametersr'   compute_default_rope_parametersr   attention_scalingregister_bufferclone)selfr%   devicerope_init_fnr$   	__class__s        d/mnt/e/genesis-system/.venv/lib/python3.12/site-packages/transformers/models/afmoe/modeling_afmoe.pyr,   zAfmoeRotaryEmbedding.__init__/   s    "("@"@$*$B$B!44[A!%!E!E>>Y&.t~~>L+7V+L($(ZeD0(..2BuU    r6   ztorch.deviceseq_lenreturnztorch.Tensorc                    | 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).
        
rope_thetahead_dimNg      ?r      dtyper6   rB   )	r0   getattrhidden_sizenum_attention_headstorcharangeint64tofloat)r%   r6   r;   basedimattention_factorr$   s          r9   r1   z4AfmoeRotaryEmbedding.compute_default_rope_parameters?   s    & %%l3fj$/c63E3EIcIc3c U\\!S!5;;?BB&X]XcXcBdgjjk
 )))r:   c                 N   | 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   AxY w)Nr   r    mpscpuF)device_typeenabledr@   rM   rA   )r$   rK   expandshaperJ   r6   
isinstancetypestrr   	transposerG   catcosr2   sinrB   )
r5   xposition_idsinv_freq_expandedposition_ids_expandedrS   freqsembr]   r^   s
             r9   forwardzAfmoeRotaryEmbedding.forward]   sR    !MM$4-8>>@GGHZHZ[\H]_acdehhijiqiqr ,QaZ 8 > > @'1!((--'E!((--[`J`ahhmmfkUC 	5&,,.1F1L1L1NNYYZ[]^_E))UEN3C'')d444C'')d444C		5 vvAGGv$cff177f&;;;	5 	5s   BFF$N)NNN)__name__
__module____qualname__rG   Tensor__annotations__r!   r,   staticmethodr   inttuplerK   r1   no_gradr   re   __classcell__r8   s   @r9   r#   r#   ,   s    llV{ V  %)+/"*d"*(* t* 
~u$	%	* *: U]]_<  <r:   r#   RMSNormc                   P     e Zd Zddeddf fdZdej                  fdZd Z xZ	S )AfmoeRMSNormepsr<   Nc                     t         |           t        j                  t	        j
                  |            | _        || _        y)z;
        AfmoeRMSNorm is equivalent to T5LayerNorm
        N)r+   r,   r   	ParameterrG   onesweightvariance_epsilon)r5   rE   ru   r8   s      r9   r,   zAfmoeRMSNorm.__init__o   s1     	ll5::k#:; #r:   c                 "   |j                   }|j                  t        j                        }|j	                  d      j                  dd      }|t        j                  || j                  z         z  }| j                  |z  j                  |      S )Nr@   rP   T)keepdim)	rB   rJ   rG   float32powmeanrsqrtrz   ry   )r5   hidden_statesinput_dtypevariances       r9   re   zAfmoeRMSNorm.forwardw   sy    #))%((7 $$Q',,R,>%Ht?T?T4T(UUm+//<<r:   c                 ^    t        | j                  j                         d| j                   S )Nz, eps=)rn   ry   rW   rz   )r5   s    r9   
extra_reprzAfmoeRMSNorm.extra_repr~   s*    ))*+6$2G2G1HIIr:   )gư>)
rg   rh   ri   rK   r,   rG   rj   re   r   rp   rq   s   @r9   rt   rt   m   s,    $ $$ $= =Jr:   rt   c                   &     e Zd Zd fd	Zd Z xZS )AfmoeMLPc                    t         |           || _        |j                  | _        ||j                  n|| _        t        j                  | j                  | j                  d      | _        t        j                  | j                  | j                  d      | _        t        j                  | j                  | j                  d      | _	        t        |j                     | _        y NFbias)r+   r,   r%   rE   intermediate_sizer   Linear	gate_projup_proj	down_projr   
hidden_actact_fn)r5   r%   r   r8   s      r9   r,   zAfmoeMLP.__init__   s    !--=N=V!9!9\m4#3#3T5K5KRWXyy!1!143I3IPUV4#9#94;K;KRWXV../r:   c                     | j                  | j                  | j                  |            | j                  |      z        }|S rf   )r   r   r   r   )r5   r_   r   s      r9   re   zAfmoeMLP.forward   s6    NN4;;t~~a/@#ADLLQRO#ST	r:   rf   )rg   rh   ri   r,   re   rp   rq   s   @r9   r   r      s    0r:   r   c                   Z     e Zd ZdZ fdZdej                  dej                  fdZ xZS )AfmoeTokenChoiceRouterz
    Token-choice top-K router for MoE routing.

    This router assigns each token to the top-K experts based on sigmoid scores, matching the released checkpoints.
    c                     t         |           || _        |j                  | _        |j
                  | _        |j                  | _        t        j                  |j                  |j
                  d      | _
        y r   )r+   r,   r%   num_experts_per_toktop_knum_expertsroute_scaler   r   rE   gater5   r%   r8   s     r9   r,   zAfmoeTokenChoiceRouter.__init__   s^    //
!--!--IIf00&2D2D5Q	r:   r   expert_biasc                    |j                   \  }}}|j                  d|      }t        j                  | j	                  |      j                  t        j                              }t        j                  ||z   | j                  d      \  }}|j                  d|      }|j                  dd      dz   }||z  }|| j                  z  }||fS )NrP   r    )krM   )rM   indexT)rM   r|   g#B;)rW   viewrG   sigmoidr   rJ   r}   topkr   gathersumr   )	r5   r   r   _
hidden_dimscoresselected_experts
top_scoresdenominators	            r9   re   zAfmoeTokenChoiceRouter.forward   s    (..1j%**2z:tyy7::5==IJ#jj+)=QRS]]q0@]A
 nnTn:UB+-
$"2"22
+++r:   )	rg   rh   ri   __doc__r,   rG   rj   re   rp   rq   s   @r9   r   r      s)    R,U\\ , ,r:   r   c                        e Zd ZdZdef fdZdej                  dej                  dej                  dej                  fdZ xZ	S )	AfmoeExpertsz
    Container holding the routed experts.

    This mirrors the Experts pattern used across other MoE models to ease checkpoint conversion.
    r%   c                     t         |           |j                  | _        |j                  | _        t        | j                        D ](  }| j                  t        ||j                               * y )N)r   )	r+   r,   r   r   r   rangeappendr   moe_intermediate_size)r5   r%   r   r8   s      r9   r,   zAfmoeExperts.__init__   s^    //
!--t''( 	ZAKK6;W;WXY	Zr:   r   r   routing_weightsr<   c                 0   |j                   \  }}}|dk(  r|j                  |d|      S |j                  d|      }|j                   d   }t        j                  |j                   d   |j
                  t        j                        j                  |      }	|j                  d      }
|j                  d      }t        j                  |
d      }|	|   }	|
|   }
||   }|j                  d|	      }t        j                  |      }t        j                  |
d      \  }}d}t        |j                         |j                               D ]'  \  }}|dk(  r||z   }||| } | |   |      }|||| |}) |j                  t        j                         |j#                  d      z  j                  |j$                        }t        j                  |      }|	j#                  d      j'                  |      }|j)                  d||       |j                  |||      S )z
        Args:
            hidden_states: (batch, seq, hidden)
            selected_experts: (batch, seq, top_k)
            routing_weights: (batch, seq, top_k)
        r   rP   rC   T)stable)return_counts)rW   	new_zerosr   rG   rH   r6   longrepeat_interleavereshapeargsortindex_select
zeros_likeunique_consecutiveziptolistrJ   r}   	unsqueezerB   	expand_asscatter_add_)r5   r   r   r   
batch_sizer;   r   hidden_states_flatr   token_indicesexpert_indicessortingdispatched_tokensexpert_outputsunique_expertscountsstart	expert_idcountendexpert_inputexpert_outputweighted_outputs
aggregatedscatter_indicess                            r9   re   zAfmoeExperts.forward   s    +8*=*='
GZa< **:q*EE*//J? &&r* $$Q'0D0DEJJ


E
" 	 *11"5)11"5--t<%g.'0)'2.;;A}M))*;<!&!9!9.X\!] #N$9$9$;V]]_ M 	Iuz%-C,U37L+DOL9M(5N5%E	 +--emm<?X?XY[?\\``anatatu%%&89
'11"5??@PQ?4DEz7J??r:   )
rg   rh   ri   r   r!   r,   rG   rj   re   rp   rq   s   @r9   r   r      sR    Z{ Z-@"\\-@=B\\-@\a\h\h-@	-@r:   r   c                   (     e Zd ZdZ fdZd Z xZS )AfmoeMoEz
    Mixture of Experts (MoE) module for AFMoE.

    This module implements a sparse MoE layer with both shared experts (always active) and
    routed experts (activated based on token-choice routing).
    c                 2   t         |           || _        t        |      | _        t        ||j                  |j                  z        | _        t        |      | _
        t        j                  t        j                  |j                        d      | _        y )NF)requires_grad)r+   r,   r%   r   routerr   r   num_shared_expertsshared_expertsr   expertsr   rw   rG   zerosr   r   r   s     r9   r,   zAfmoeMoE.__init__   sp    ,V4&vv/K/KfNgNg/gh#F+<<F4F4F(GW\]r:   c                    |j                   \  }}}|j                  d|      }| j                  || j                        \  }}|j                  ||| j                  j
                        }|j                  ||| j                  j
                        }| j                  |      j                  |||      }| j                  |||      }	||	z   S )NrP   )rW   r   r   r   r%   r   r   r   )
r5   r   r   r;   r   r   r   r   shared_outputrouted_outputs
             r9   re   zAfmoeMoE.forward   s    *7*=*='
GZ*//J? (,{{=$BRBR'S$
$__Z$++:Y:YZ
+00WdkkFeFef ++,>?DDZQXZde]4DjQ},,r:   )rg   rh   ri   r   r,   re   rp   rq   s   @r9   r   r      s    ^-r:   r   c                     | 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..NrP   r@   rU   )rW   rG   r\   )r_   x1x2s      r9   rotate_halfr     sZ    	
3"!''"+"""	#B	
3q ""	#B99rc2YB''r:   rotary_pos_embc                     |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.
    )r   r   )qr   r]   r^   unsqueeze_dimq_embedk_embeds          r9   apply_rotary_pos_embr     sY    & --
&C
--
&C3w;q>C/0G3w;q>C/0GGr:   r   n_repr<   c                     | j                   \  }}}}|dk(  r| S | dddddddddf   j                  |||||      } | j                  |||z  ||      S )z
    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)rW   rV   r   )r   r   batchnum_key_value_headsslenr?   s         r9   	repeat_kvr   -  so    
 2?1D1D.Ehz!!Qa"23::5BUW\^bdlmM  (;e(CT8TTr:   modulequerykeyvalueattention_maskscalingdropoutkwargsc                    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   rP   )rM   rB   )ptrainingr    )r   num_key_value_groupsrG   matmulr[   r   
functionalsoftmaxr}   rJ   rB   r   r   
contiguous)r   r   r   r   r   r   r   r   
key_statesvalue_statesattn_weightsattn_outputs               r9   eager_attention_forwardr  9  s     3 ; ;<JUF$?$?@L<<z';';Aq'ABWLL!#n4==((2U]](SVVW\WbWbcL==((6??([L,,|\:K''1-88:K$$r:   c                       e Zd ZdZdedef 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j                  dz  dee   de	ej                  ej                  f   fdZ xZS )AfmoeAttentionaJ  
    Multi-headed attention module with optional sliding window and gating.

    This attention mechanism supports both full attention and sliding window attention,
    and includes Q/K normalization and gating of the output. It inherits from [`LlamaAttention`] to minimize the amount
    of custom logic we need to maintain.
    r%   	layer_idxc                    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                        | _        |j(                  |   dk(  | _        | j*                  r|j,                  nd | _        t/        | j                  |j0                        | _        t/        | j                  |j0                        | _        t        j                  |j
                  |j                  | j                  z  d      | _        y )Nr?   g      Tr   sliding_attentionru   F)r+   r,   r%   r
  rD   rE   rF   r?   r   r   r   attention_dropout	is_causalr   r   attention_biasq_projk_projv_projo_projlayer_typesis_local_attentionsliding_windowrt   rms_norm_epsq_normk_normr   r5   r%   r
  r8   s      r9   r,   zAfmoeAttention.__init__\  s   "
F4F4F&JdJd4de$*$>$>&B\B\$\!}}d*!'!9!9ii : :T]] JQWQfQf
 ii : :T]] JQWQfQf
 ii : :T]] JQWQfQf
 ii&&68J8JQWQfQf

 #)"4"4Y"?CV"V7;7N7Nf33TX"4==f6I6IJ"4==f6I6IJ6#5#5v7Q7QTXTaTa7ahmnr:   Nr   position_embeddingsr   past_key_valuecache_positionr   r<   c                    |j                   d d }g |d| j                  }| j                  |      j                  |      }	| j	                  |      j                  |      }
| j                  |      j                  |      }| j                  |      }| j                  |	      j                  dd      }	| j                  |
      j                  dd      }
|j                  dd      }| j                  r|\  }}t        |	|
||      \  }	}
|%d|i}|j                  |
|| j                  |      \  }
}t        j                  | j                   j"                  t$              } || |	|
|f|| j&                  sdn| j(                  | j*                  | j,                  d|\  }} |j                  g |d j/                         }|t1        j2                  |      z  }| j5                  |      }||fS )NrP   r    r@   r          )r   r   r   r  )rW   r?   r  r   r  r  r   r  r[   r  r  r   updater
  r   get_interfacer%   _attn_implementationr  r   r  r   r  r  rG   r   r  )r5   r   r  r   r  r  r   input_shapehidden_shapequery_statesr  r  gate_statesr]   r^   cache_kwargsattention_interfaceoutputr  r  s                       r9   re   zAfmoeAttention.forward{  s    $))#2.88b8$--8{{=166|D[[/44\B
{{=166|Dnn]3{{<0::1a@[[,66q!<
#--a3""*HC';L*VY[^'_$L*%,n=L'5'<'<ZW[WeWegs't$J(?(M(MKK,,.E)
  3	
 

 *#}}C$2H2HLL..
 
 
 
 .k.2.99;%--44kk&)L((r:   )NN)rg   rh   ri   r   r!   rm   r,   rG   rj   rn   r	   
LongTensorr   r   re   rp   rq   s   @r9   r	  r	  R  s    o{ os oH (,260)||0) #5<<#=>0) t+	0)
 0) ((4/0) +,0) 
u||U\\)	*0)r:   r	  c                   &    e Zd ZdZdedef 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j                  dz  deej                  ej                  f   dz  dee   dej                  fdZ xZS )AfmoeDecoderLayerz
    AFMoE decoder layer with dual normalization.

    This layer applies self-attention followed by either a dense MLP or MoE block,
    with dual normalization (pre and post) around each component.
    r%   r
  c                 P   t         |           |j                  | _        || _        t	        ||      | _        |j                  |   | _        t        |j                  |j                        | _
        t        |j                  |j                        | _        t        |j                  |j                        | _        t        |j                  |j                        | _        ||j                  k\  | _        | j                  rt!        |      | _        y t%        |      | _        y )N)r%   r
  r  )r+   r,   rE   r
  r	  	self_attnr  attention_typert   r  input_layernormpost_attention_layernormpre_mlp_layernormpost_mlp_layernormnum_dense_layersmoe_enabledr   mlpr   r  s      r9   r,   zAfmoeDecoderLayer.__init__  s    !--"'vK$00;  ,F,>,>FDWDWX(4V5G5GVM`M`(a% ".f.@.@fFYFY!Z".v/A/AvGZGZ"[ %(?(??'DH'DHr:   Nr   r   r`   r  	use_cacher  r  r   r<   c                    |}	| j                  |      } | j                  d|||||||d|\  }}
| j                  |      }|	|z   }|}	| j                  |      }| j	                  |      }| j                  |      }|	|z   }|S )N)r   r   r`   r  r8  r  r   )r1  r/  r2  r3  r7  r4  )r5   r   r   r`   r  r8  r  r  r   residualr   s              r9   re   zAfmoeDecoderLayer.forward  s     ! ,,];)4>> 	
')%)) 3	
 	
q 55mD =0 !..}=///> =0r:   )NNNNNN)rg   rh   ri   r   r!   rm   r,   rG   rj   r+  r	   boolrn   r   r   FloatTensorre   rp   rq   s   @r9   r-  r-    s    ({ (s (4 /304'+!%26HL#||# t+# &&-	#
 # $;# ((4/# #5<<#=>E# +,# 
		#r:   r-  c                   d     e Zd ZU dZeed<   dZdgZdgZe	e
dZg dZdZdZdZdZdZ fd	Z xZS )
AfmoePreTrainedModelz
    An abstract class to handle weights initialization and a simple interface for downloading and loading pretrained
    models.
    r%   modelr-  past_key_values)r   
attentions)r1  r2  r3  r4  r  r  normr   Tc                     t         |   |       t        |t              r*t	        j
                  |j                  j                         yt        |t              r t	        j
                  |j                         yy)zInitialize the weightsN)
r+   _init_weightsrX   r   initzeros_r   ry   r   r   )r5   r   r8   s     r9   rE  z"AfmoePreTrainedModel._init_weights  sR    f%f45KK**+)KK**+ *r:   )rg   rh   ri   r   r!   rk   base_model_prefix_no_split_modules_skip_keys_device_placementr-  r	  _can_record_outputs_keep_in_fp32_modules_supports_sdpa_supports_flash_attn_supports_flex_attn_supports_attention_backendsupports_gradient_checkpointingrE  rp   rq   s   @r9   r?  r?    sg    
 ,-#4"5*$	 N"&&*#, ,r:   r?  c                       e Zd ZdZdef fdZeee	 	 	 	 	 	 	 dde	j                  dz  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ez  fd                     Z xZS )
AfmoeModelz
    Transformer decoder consisting of *config.num_hidden_layers* layers. Each layer is a [`AfmoeDecoderLayer`]

    Args:
        config: AfmoeConfig
    r%   c           	         t         |   |       |j                  | _        |j                  | _        t        j                  |j                  |j                  | j                        | _        t        j                  t        |j                        D cg c]  }t        ||       c}      | _        t        |j                  |j                        | _        t#        |      | _        d| _        | j)                          y c c}w )Nr  r%   F)r+   r,   pad_token_idpadding_idx
vocab_sizer   	EmbeddingrE   embed_tokens
ModuleListr   num_hidden_layersr-  layersrt   r  rC  r#   
rotary_embgradient_checkpointing	post_initr  s      r9   r,   zAfmoeModel.__init__#  s     !.. ++LL):):F<N<NPTP`P`ammCHIaIaCbcivy1c
 !!3!39L9LM	.f=&+# ds   DN	input_idsr   inputs_embedsr`   rA  r  r8  r   r<   c                    |d u |d uz  rt        d      |r|t        | j                        }|| j                  |      }|F||j	                         nd}	t        j                  |	|	|j                  d   z   |j                        }||j                  d      }t        |x}
t              s)| j                  ||||d}t        di |t        di |d}
|}| j                  j                  r|| j                  j                  dz  z  }| j!                  ||      }| j"                  D ]  } ||f|
|j$                     |||||d	|}! | j'                  |      }t)        ||r|
      S d 
      S )Nz:You must specify exactly one of input_ids or inputs_embedsrU  r   r    )r6   )r%   rb  r   r  rA  )full_attentionr  g      ?)r   r`   r  r8  r  r  )last_hidden_staterA  r:  )
ValueErrorr
   r%   rZ  get_seq_lengthrG   rH   rW   r6   r   rX   dictr   r   mup_enabledrE   r^  r]  r0  rC  r   )r5   ra  r   rb  r`   rA  r  r8  r   past_seen_tokenscausal_mask_mappingmask_kwargsr   r  decoder_layers                  r9   re   zAfmoeModel.forward2  s    -t";<YZZ0*$++>O  --i8M!CRC^==?de"\\  =#6#6q#99$++N
 )33A6L ?-F++!."0"0#2K #5"C{"C%F%U%U#
 & ;;"")T[[-D-Dc-IJM"oom\J![[ 
	M)	2=3O3OP).#-$7	 	M
	 		-0%+/8O
 	
>B
 	
r:   )NNNNNNN)rg   rh   ri   r   r!   r,   r   r   r   rG   r+  rj   r=  r	   r<  r   r   rn   r   re   rp   rq   s   @r9   rS  rS    s    {   .2.22604(,26!%D
##d*D
 t+D
 ((4/	D

 &&-D
 D
 ((4/D
 $;D
 +,D
 
'	'D
    D
r:   rS  c                   b    e Zd ZddiZddiZddgdgfiZ 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	j                  dz  dee	j                  z  dee   defd              Z xZS )AfmoeForCausalLMzlm_head.weightzmodel.embed_tokens.weightlm_headcolwise_gather_outputr   logitsc                     t         |   |       t        |      | _        |j                  | _        t        j                  |j                  |j                  d      | _        | j                          y r   )
r+   r,   rS  r@  rX  r   r   rE   rp  r`  r   s     r9   r,   zAfmoeForCausalLM.__init__  sS     '
 ++yy!3!3V5F5FUSr:   Nra  r   r`   rA  rb  labelsr8  r  logits_to_keepr   r<   c
                 z    | j                   d|||||||d|
}|j                  }t        |	t              rt	        |	 d      n|	}| j                  |dd|ddf         }d}|* | j                  d||| j                  j                  d|
}t        |||j                  |j                  |j                        S )a  
        Example:

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

        >>> model = AfmoeForCausalLM.from_pretrained("meta-afmoe/Afmoe-2-7b-hf")
        >>> tokenizer = AutoTokenizer.from_pretrained("meta-afmoe/Afmoe-2-7b-hf")

        >>> 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."
        ```)ra  r   r`   rA  rb  r8  r  N)rr  rt  rX  )lossrr  rA  r   rB  r:  )r@  re  rX   rm   slicerp  loss_functionr%   rX  r   rA  r   rB  )r5   ra  r   r`   rA  rb  rt  r8  r  ru  r   outputsr   slice_indicesrr  rw  s                   r9   re   zAfmoeForCausalLM.forward  s    @ ,64:: 	,
)%+')	,
 	,
  118B>SV8W~ot4]kmA}a,?@A%4%%pVFt{{OeOepiopD%#33!//))
 	
r:   )	NNNNNNNNr   )rg   rh   ri   _tied_weights_keys_tp_plan_pp_planr,   r   r   rG   r+  rj   r	   r=  r<  rm   r   r   r   re   rp   rq   s   @r9   ro  ro  |  s/   *,GH23H_-z:;H  .2.204(,26*.!%26-.8
##d*8
 t+8
 &&-	8

 8
 ((4/8
   4'8
 $;8
 ((4/8
 ell*8
 +,8
 
 8
  8
r:   ro  )ro  rS  r?  )r    )r   )Ecollections.abcr   typingr   rG   r    r   rF  activationsr   cache_utilsr	   r
   
generationr   integrationsr   r   r   masking_utilsr   r   modeling_layersr   modeling_outputsr   r   r   modeling_rope_utilsr   r   modeling_utilsr   r   processing_utilsr   utilsr   r   r   utils.genericr   r   utils.output_capturingr   configuration_afmoer!   Moduler#   rt   r   r   r[  r   r   r   r   rj   rm   r   rK   r  r	  r-  r?  rS  ro  __all__r:  r:   r9   <module>r     s  * %    & ! . ) f f R 9 g g K F & I I G 5 ,><299 ><B Y'J299 J (J(ryy  ,RYY ,:;@2== ;@|-ryy ->( *+ ,2	UU\\ 	U# 	U%,, 	U& %II%<<% 
% <<	%
 LL4'% % % '(%2 )*X)RYY X) +X)vB2 BJ$,? $,N ^
% ^
 ^
B F
+_ F
 F
R Er:   