
    rh
                    0   d Z ddlZddlmZmZ ddlZddl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 ddlmZmZmZmZmZmZ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)  e'jT                  e+      Z, G d dejZ                        Z. G d dejZ                        Z/ G d dejZ                        Z0 G d dejZ                        Z1 G d dejZ                        Z2 G d dejZ                        Z3 G d dejZ                        Z4 G d d e      Z5 G d! d"ejZ                        Z6 G d# d$ejZ                        Z7e& G d% d&e              Z8 e&d'(       G d) d*e8             Z9 e&d+(       G d, d-e8e             Z:e& G d. d/e8             Z; G d0 d1ejZ                        Z< e&d2(       G d3 d4e8             Z=e& G d5 d6e8             Z>e& G d7 d8e8             Z? G d9 d:ejZ                        Z@e& G d; d<e8             ZAd?d=ZBg d>ZCy)@zPyTorch X-MOD model.    N)OptionalUnion)nn)BCEWithLogitsLossCrossEntropyLossMSELoss   )ACT2FNgelu)CacheEncoderDecoderCache)GenerationMixin)GradientCheckpointingLayer))BaseModelOutputWithPastAndCrossAttentions,BaseModelOutputWithPoolingAndCrossAttentions!CausalLMOutputWithCrossAttentionsMaskedLMOutputMultipleChoiceModelOutputQuestionAnsweringModelOutputSequenceClassifierOutputTokenClassifierOutput)PreTrainedModel)apply_chunking_to_forward find_pruneable_heads_and_indicesprune_linear_layer)auto_docstringlogging   )
XmodConfigc                   2     e Zd ZdZ fdZ	 ddZd Z xZS )XmodEmbeddingszV
    Same as BertEmbeddings with a tiny tweak for positional embeddings indexing.
    c                    t         |           t        j                  |j                  |j
                  |j                        | _        t        j                  |j                  |j
                        | _	        t        j                  |j                  |j
                        | _        t        j                  |j
                  |j                        | _        t        j                  |j                        | _        t#        |dd      | _        | j'                  dt)        j*                  |j                        j-                  d      d       | j'                  d	t)        j.                  | j0                  j3                         t(        j4                  
      d       |j                  | _        t        j                  |j                  |j
                  | j6                        | _	        y )N)padding_idxepsposition_embedding_typeabsoluteposition_ids)r   F)
persistenttoken_type_idsdtype)super__init__r   	Embedding
vocab_sizehidden_sizepad_token_idword_embeddingsmax_position_embeddingsposition_embeddingstype_vocab_sizetoken_type_embeddings	LayerNormlayer_norm_epsDropouthidden_dropout_probdropoutgetattrr&   register_buffertorcharangeexpandzerosr(   sizelongr#   selfconfig	__class__s     y/var/www/html/ai-insurance-compliance-backend/venv/lib/python3.12/site-packages/transformers/models/xmod/modeling_xmod.pyr/   zXmodEmbeddings.__init__7   si   !||F,=,=v?Q?Q_e_r_rs#%<<0N0NPVPbPb#c %'\\&2H2H&J\J\%]" f&8&8f>S>STzz&"<"<='.v7PR\']$ELL)G)GHOOPWXej 	 	
 	ekk$*;*;*@*@*B%**Ubg 	 	

 "..#%<<**F,>,>DL\L\$
     c                    |+|t        || j                  |      }n| j                  |      }||j                         }n|j                         d d }|d   }|st	        | d      r-| j
                  d d d |f   }|j                  |d   |      }	|	}n:t        j                  |t        j                  | j                  j                        }|| j                  |      }| j                  |      }
||
z   }| j                  dk(  r| j                  |      }||z  }| j!                  |      }| j#                  |      }|S )Nr)   r   r+   r   r-   devicer'   )"create_position_ids_from_input_idsr#   &create_position_ids_from_inputs_embedsrD   hasattrr+   rB   r@   rC   rE   r(   rN   r4   r8   r&   r6   r9   r=   )rG   	input_idsr+   r(   inputs_embedspast_key_values_lengthinput_shape
seq_lengthbuffered_token_type_ids buffered_token_type_ids_expandedr8   
embeddingsr6   s                rJ   forwardzXmodEmbeddings.forwardP   sR    $A)TM]M]_uv#JJ=Y #..*K',,.s3K ^

 !t-.*.*=*=a*n*M'3J3Q3QR]^_R`bl3m0!A!&[

SWSdSdSkSk!l  00;M $ : :> J"%::
'':5"&":":<"H--J^^J/
\\*-
rK   c                    |j                         dd }|d   }t        j                  | j                  dz   || j                  z   dz   t        j                  |j
                        }|j                  d      j                  |      S )z
        We are provided embeddings directly. We cannot infer which are padded so just generate sequential position ids.

        Args:
            inputs_embeds: torch.Tensor

        Returns: torch.Tensor
        Nr)   r   rM   r   )rD   r@   rA   r#   rE   rN   	unsqueezerB   )rG   rS   rU   sequence_lengthr(   s        rJ   rP   z5XmodEmbeddings.create_position_ids_from_inputs_embedsx   s     $((*3B/%a.||q /D4D4D"Dq"HPUPZPZcpcwcw
 %%a(//<<rK   )NNNNr   )__name__
__module____qualname____doc__r/   rZ   rP   __classcell__rI   s   @rJ   r!   r!   1   s    

4 rs&P=rK   r!   c                        e Zd Zd fd	Z	 	 	 	 	 	 ddej
                  deej                     deej                     deej                     dee   dee	   deej
                     d	e
ej
                     fd
Z xZS )XmodSelfAttentionc                    t         |           |j                  |j                  z  dk7  r2t	        |d      s&t        d|j                   d|j                   d      |j                  | _        t        |j                  |j                  z        | _        | j                  | j                  z  | _        t        j                  |j                  | j                        | _        t        j                  |j                  | j                        | _        t        j                  |j                  | j                        | _        t        j                  |j                        | _        |xs t#        |dd      | _        | j$                  dk(  s| j$                  d	k(  rF|j&                  | _        t        j(                  d
|j&                  z  dz
  | j                        | _        |j,                  | _        || _        y )Nr   embedding_sizezThe hidden size (z6) is not a multiple of the number of attention heads ()r&   r'   relative_keyrelative_key_query   r   )r.   r/   r2   num_attention_headsrQ   
ValueErrorintattention_head_sizeall_head_sizer   Linearquerykeyvaluer;   attention_probs_dropout_probr=   r>   r&   r5   r0   distance_embedding
is_decoder	layer_idxrG   rH   r&   rx   rI   s       rJ   r/   zXmodSelfAttention.__init__   s    : ::a?PVXhHi#F$6$6#7 8 445Q8 
 $*#=#= #&v'9'9F<V<V'V#W !558P8PPYYv1143E3EF
99V//1C1CDYYv1143E3EF
zz&"E"EF'> (
'-zC
$ ''>9T=Y=Y]q=q+1+I+ID(&(ll1v7U7U3UXY3Y[_[s[s&tD# ++"rK   hidden_statesattention_mask	head_maskencoder_hidden_statespast_key_valueoutput_attentionscache_positionreturnc                 	   |j                   \  }}	}
| j                  |      }|j                  |d| j                  | j                        j                  dd      }|d u}|St        |t              rA|j                  j                  | j                        }|r|j                  }n|j                  }n|}|r|n|}|rK|IrGj                  | j                     j                  }|j                  | j                     j                  }n| j!                  |      }|j                  |d| j                  | j                        j                  dd      }| j#                  |      }|j                  |d| j                  | j                        j                  dd      }|D|s|nd }j%                  ||| j                  d|i      \  }}|rd|j                  | j                  <   t'        j(                  ||j                  dd            }| j*                  dk(  s| j*                  dk(  r|j                   d   |j                   d   }}|Dt'        j,                  |dz
  t&        j.                  |j0                  	      j                  dd      }n@t'        j2                  |t&        j.                  |j0                  	      j                  dd      }t'        j2                  |t&        j.                  |j0                  	      j                  dd      }||z
  }| j5                  || j6                  z   dz
        }|j9                  |j:                  
      }| j*                  dk(  rt'        j<                  d||      }||z   }nE| j*                  dk(  r6t'        j<                  d||      }t'        j<                  d||      }||z   |z   }|t?        j@                  | j                        z  }|||z   }tB        jD                  jG                  |d      }| jI                  |      }|||z  }t'        j(                  ||      }|jK                  dddd      jM                         }|jO                         d d | jP                  fz   }|j                  |      }||fS )Nr)   r   rk   r   Tri   rj   rM   r,   zbhld,lrd->bhlrzbhrd,lrd->bhlrdimr   r	   ))shaperr   viewrl   ro   	transpose
isinstancer   
is_updatedgetrx   cross_attention_cacheself_attention_cachelayerskeysvaluesrs   rt   updater@   matmulr&   tensorrE   rN   rA   rv   r5   tor-   einsummathsqrtr   
functionalsoftmaxr=   permute
contiguousrD   rp   )rG   rz   r{   r|   r}   r~   r   r   
batch_sizerV   _query_layeris_cross_attentionr   curr_past_key_valuecurrent_states	key_layervalue_layerattention_scoresquery_length
key_lengthposition_ids_lposition_ids_rdistancepositional_embeddingrelative_position_scoresrelative_position_scores_queryrelative_position_scores_keyattention_probscontext_layernew_context_layer_shapes                                  rJ   rZ   zXmodSelfAttention.forward   sN    %2$7$7!
Jjj/!&&z2t7O7OQUQiQijttq
 3$>%.*=>+66::4>>J
%*8*N*N'*8*M*M'&4#2D.-."<+224>>BGGI-44T^^DKKK0I!z2t7O7OQUQiQijtt1I **^4K%**B 8 8$:R:Ri1o  )7It)<)C)C{DNN=M~<^*&	; &@DN--dnn= !<<Y5H5HR5PQ''>9T=Y=Y]q=q'2'8'8';Y__Q=O*L)!&j1nEJJWdWkWk!l!q!q" "'l%**UbUiUi!j!o!oprtu!v"\\*EJJ}OcOcdiijkmopN%6H#'#:#:8dFbFb;bef;f#g #7#:#:ARAR#:#S ++~=+0<<8H+Wk+l(#36N#N --1EE16>NP[]q1r./4||<LiYm/n,#36T#TWs#s +dii8P8P.QQ%/.@ --//0@b/I ,,7  -	9O_kB%--aAq9DDF"/"4"4"6s";t?Q?Q>S"S%**+BCo--rK   NNNNNNFN)r^   r_   r`   r/   r@   Tensorr   FloatTensorr   booltuplerZ   rb   rc   s   @rJ   re   re      s    #< 7;15=A*.,115d.||d. !!2!23d. E--.	d.
  ((9(9:d. !d. $D>d. !.d. 
u||	d.rK   re   c                   n     e Zd Z fdZdej
                  dej
                  dej
                  fdZ xZS )XmodSelfOutputc                 (   t         |           t        j                  |j                  |j                        | _        t        j                  |j                  |j                        | _        t        j                  |j                        | _
        y Nr$   )r.   r/   r   rq   r2   denser9   r:   r;   r<   r=   rF   s     rJ   r/   zXmodSelfOutput.__init__  s`    YYv1163E3EF
f&8&8f>S>STzz&"<"<=rK   rz   input_tensorr   c                 T    | j                  |      }| j                  |      }||z   }|S N)r   r=   )rG   rz   r   s      rJ   rZ   zXmodSelfOutput.forward  s.    

=1]3%4rK   r^   r_   r`   r/   r@   r   rZ   rb   rc   s   @rJ   r   r     s1    >U\\  RWR^R^ rK   r   c                        e Zd Zd fd	Zd Z	 	 	 	 	 	 ddej                  deej                     deej                     deej                     dee	   dee
   d	eej                     d
eej                     fdZ xZS )XmodAttentionc                     t         |           t        |||      | _        t	        |      | _        t               | _        |j                  | _        y )Nr&   rx   )	r.   r/   re   rG   r   outputsetpruned_headspre_normry   s       rJ   r/   zXmodAttention.__init__  sA    %fF]irs	$V,ErK   c                 >   t        |      dk(  ry t        || j                  j                  | j                  j                  | j
                        \  }}t        | j                  j                  |      | j                  _        t        | j                  j                  |      | j                  _        t        | j                  j                  |      | j                  _	        t        | j                  j                  |d      | j                  _        | j                  j                  t        |      z
  | j                  _        | j                  j                  | j                  j                  z  | j                  _        | j
                  j                  |      | _        y )Nr   r   r   )lenr   rG   rl   ro   r   r   rr   rs   rt   r   r   rp   union)rG   headsindexs      rJ   prune_headszXmodAttention.prune_heads&  s   u:?749900$))2O2OQUQbQb
u
 -TYY__eD		*499==%@		,TYY__eD		.t{{/@/@%QO )-		(E(EE
(R		%"&))"?"?$))B_B_"_		 --33E:rK   rz   r{   r|   r}   r~   r   r   r   c           	         |}| j                   r| j                  j                  |      }| j                  |||||||      }	| j                  |	d   |      }
| j                   s| j                  j                  |
      }
|
f|	dd  z   }|S )Nr   r   )r   r   r9   rG   )rG   rz   r{   r|   r}   r~   r   r   residualself_outputsattention_outputoutputss               rJ   rZ   zXmodAttention.forward8  s     !== KK11-@Myy!
  ;;|AA}}#{{445EF#%QR(88rK   r   r   )r^   r_   r`   r/   r   r@   r   r   r   r   r   r   rZ   rb   rc   s   @rJ   r   r     s    (;* 7;15=A*.,115|| !!2!23 E--.	
  ((9(9: ! $D> !. 
u||	rK   r   c                   V     e Zd Z fdZdej
                  dej
                  fdZ xZS )XmodIntermediatec                    t         |           t        j                  |j                  |j
                        | _        t        |j                  t              rt        |j                     | _        y |j                  | _        y r   )r.   r/   r   rq   r2   intermediate_sizer   r   
hidden_actstrr
   intermediate_act_fnrF   s     rJ   r/   zXmodIntermediate.__init__W  s]    YYv1163K3KL
f''-'-f.?.?'@D$'-'8'8D$rK   rz   r   c                 J    | j                  |      }| j                  |      }|S r   )r   r   rG   rz   s     rJ   rZ   zXmodIntermediate.forward_  s&    

=100?rK   r   rc   s   @rJ   r   r   V  s#    9U\\ ell rK   r   c                   V     e Zd Z fdZdej
                  dej
                  fdZ xZS )XmodAdapterc                    t         |           |j                  |j                  z  | _        t        j                  |j                  | j                        | _        t        j                  | j                  |j                        | _        t        |j                  t              rt        |j                     | _        y |j                  | _        y r   )r.   r/   r2   adapter_reduction_factorbottleneck_sizer   rq   dense1dense2r   r   r   r
   adapter_act_fnrF   s     rJ   r/   zXmodAdapter.__init__f  s    %11V5T5TTii 2 2D4H4HIii 4 4f6H6HIf''-"():):";D"("3"3DrK   rz   r   c                 l    | j                  |      }| j                  |      }| j                  |      }|S r   )r   r   r   r   s     rJ   rZ   zXmodAdapter.forwardp  s4    M2++M:M2rK   r   rc   s   @rJ   r   r   e  s#    4U\\ ell rK   r   c                        e Zd Z fdZdej
                  dej
                  dej
                  dej
                  fdZdej
                  dej
                  fdZ xZS )
XmodOutputc                    t         |           t        j                  |j                  |j
                        | _        t        j                  |j
                  |j                        | _        |j                  | _	        t        j                  |j                        | _        |j                  r1t        j                  |j
                  |j                        | _        nd | _        |j                  | _        t        j                  i       | _        |j"                  D ]#  }t%        |      | j                   t'        |      <   % y r   )r.   r/   r   rq   r   r2   r   r9   r:   ln_before_adapterr;   r<   r=   adapter_layer_normadapter_reuse_layer_norm
ModuleDictadapter_modules	languagesr   r   )rG   rH   languagerI   s      rJ   r/   zXmodOutput.__init__x  s    YYv779K9KL
f&8&8f>S>ST!'!9!9zz&"<"<=$$&(ll63E3E6K`K`&aD#&*D#(.(G(G%!}}R0(( 	FH2=f2ED  X/	FrK   rz   r   lang_idsr   c                 x    | j                  |      }| j                  |      }||z   }| j                  ||      }|S r   )r   r=   lang_adapter)rG   rz   r   r   s       rJ   rZ   zXmodOutput.forward  s@    

=1]3%4))(MBrK   c                    t        j                  |d      \  }}| j                  s|}| j                  | j                  |      }n| j                  r| j                  |      }| j                  r|}t        j                  ||j                         d      }g }t        t        ||            D ]i  \  }\  }}	t        | j                  j                               t        |j                                  }
|j                   | j                  |
   |	             k t        j                   |d      }| j#                  |      }|z  }|S )NT)return_countsr   )r@   unique_consecutiver   r   r   r9   splittolist	enumerateziplistr   r   rn   itemappendcatr=   )rG   r   rz   lang_lengthsr   split_hidden_stateslang_wise_outputsilang_idsplit_hidden_statelangs              rJ   r   zXmodOutput.lang_adapter  s1   !&!9!9(RV!W,%%$H"". 33MBM** NN=9M!!$H#kk-9L9L9NPQR09#hH[:\0] 	U,A,+,,1134S5HID$$%?T%9%9$%?@R%ST	U 		"3Q7]3!rK   )	r^   r_   r`   r/   r@   r   rZ   r   rb   rc   s   @rJ   r   r   w  s[    FU\\  Y^YeYe jojvjv U\\ %,, rK   r   c                   6    e Zd Zd fd	Z	 	 	 	 	 	 	 ddej
                  dej
                  deej                     deej                     deej                     deej                     dee   d	ee	   d
eej
                     de
ej
                     fdZd Z xZS )	XmodLayerc                    t         |           |j                  | _        d| _        t	        ||      | _        |j                  | _        |j                  | _        | j                  r-| j                  st        |  d      t	        |d|      | _	        t        |      | _        t        |      | _        |j                  | _        y )Nr   rx   z> should be used as a decoder model if cross attention is addedr'   r   )r.   r/   chunk_size_feed_forwardseq_len_dimr   	attentionrw   add_cross_attentionrm   crossattentionr   intermediater   r   r   )rG   rH   rx   rI   s      rJ   r/   zXmodLayer.__init__  s    '-'E'E$&vC ++#)#=#= ##?? D6)g!hii"/PZfo"pD,V4 (rK   rz   r   r{   r|   r}   encoder_attention_maskr~   r   r   r   c
           	      
   | j                  ||||||	      }
|
d   }|
dd  }| j                  rB|@t        | d      st        d|  d      | j	                  |||||||	      }|d   }||dd  z   }|}| j
                  r| j                  j                  |      }t        | j                  | j                  | j                  |      }| j                  |||      }| j
                  s| j                  j                  |      }|f|z   S )N)r{   r|   r   r~   r   r   r   r  z'If `encoder_hidden_states` are passed, z` has to be instantiated with cross-attention layers by setting `config.add_cross_attention=True`)r{   r|   r}   r~   r   r   )r
  rw   rQ   rm   r  r   r   r9   r   feed_forward_chunkr  r	  )rG   rz   r   r{   r|   r}   r  r~   r   r   self_attention_outputsr   r   cross_attention_outputsr   intermediate_outputlayer_outputs                    rJ   rZ   zXmodLayer.forward  sP    "&)/)) "0 "
 2!4(,??4@4!12 =dV DD D 
 '+&9&9 5#&;-"3- ': '#  7q9 7 ;;G#==#{{445EF7##((	
 {{#6(K}};;00>L((rK   c                 $    | j                  |      S r   )r  )rG   r   s     rJ   r  zXmodLayer.feed_forward_chunk  s      !122rK   r   )NNNNNFN)r^   r_   r`   r/   r@   r   r   r   r   r   r   rZ   r  rb   rc   s   @rJ   r  r    s    (& 7;15=A>B*.,1156)||6) ,,6) !!2!23	6)
 E--.6)  ((9(9:6) !)):): ;6) !6) $D>6) !.6) 
u||	6)p3rK   r  c                   |    e Zd Z fdZ	 	 	 	 	 	 	 	 	 	 ddej
                  dej
                  deej                     deej                     deej                     deej                     deeeej                           d	ee	   d
ee	   dee	   dee	   deej
                     de
eej
                     ef   fdZ xZS )XmodEncoderc           	      n   t         |           || _        t        j                  t        |j                        D cg c]  }t        ||       c}      | _        |j                  | _
        | j                  r0t        j                  |j                  |j                        | _        d| _        y c c}w )Nr  r$   F)r.   r/   rH   r   
ModuleListrangenum_hidden_layersr  layerr   is_pre_normr9   r2   r:   gradient_checkpointing)rG   rH   r   rI   s      rJ   r/   zXmodEncoder.__init__  s    ]]ERXRjRjLk#lqIf$B#lm
!??\\&*<*<&BWBWXDN&+#	 $ms   B2rz   r   r{   r|   r}   r  past_key_values	use_cacher   output_hidden_statesreturn_dictr   r   c                    | j                   r%| j                  r|rt        j                  d       d}d}|r<t	        |t
              s,t        j                  d       d}t        j                  |      }|
rdnd }|	rdnd }|	r| j                  j                  rdnd }t        | j                        D ]W  \  }}|
r||fz   }|||   nd } |||||||||	|	      }|d   }|	s/||d   fz   }| j                  j                  sO||d   fz   }Y | j                  r| j                  |      }|
r||fz   }|r|j                         }|st        d	 |||||fD              S t!        |||||
      S )NzZ`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...FzPassing a tuple of `past_key_values` is deprecated and will be removed in Transformers v4.58.0. You should pass an instance of `EncoderDecoderCache` instead, e.g. `past_key_values=EncoderDecoderCache.from_legacy_cache(past_key_values)`.T r   r   rk   c              3   $   K   | ]  }|| 
 y wr   r$  ).0vs     rJ   	<genexpr>z&XmodEncoder.forward.<locals>.<genexpr>E  s      
 = 
s   )last_hidden_stater  rz   
attentionscross_attentions)r  trainingloggerwarning_oncer   r   r   from_legacy_cacherH   r  r   r  r  r9   to_legacy_cacher   r   )rG   rz   r   r{   r|   r}   r  r  r   r   r!  r"  r   return_legacy_cacheall_hidden_statesall_self_attentionsall_cross_attentionsr   layer_modulelayer_head_masklayer_outputss                        rJ   rZ   zXmodEncoder.forward  s    &&4==##p "	#Z?\
 #'1CCOTO"6BD$5b4%64;;;Z;Zr`d(4 	VOA|#$58H$H!.7.CilO(%&!
M *!,M &9]1=M<O&O#;;22+?=QRCSBU+U(-	V0  NN=9M 1]4D D-==?O 
 "#%'(
 
 
 9+++*1
 	
rK   )
NNNNNNFFTN)r^   r_   r`   r/   r@   r   r   r   r   r   r   r   rZ   rb   rc   s   @rJ   r  r    s1   , 7;15=A>BEI$(,1/4&*15W
||W
 ,,W
 !!2!23	W

 E--.W
  ((9(9:W
 !)):): ;W
 "%e.?.?(@"ABW
 D>W
 $D>W
 'tnW
 d^W
 !.W
 
uU\\"$MM	NW
rK   r  c                   V     e Zd Z fdZdej
                  dej
                  fdZ xZS )
XmodPoolerc                     t         |           t        j                  |j                  |j                        | _        t        j                         | _        y r   )r.   r/   r   rq   r2   r   Tanh
activationrF   s     rJ   r/   zXmodPooler.__init__[  s9    YYv1163E3EF
'')rK   rz   r   c                 \    |d d df   }| j                  |      }| j                  |      }|S Nr   )r   r<  )rG   rz   first_token_tensorpooled_outputs       rJ   rZ   zXmodPooler.forward`  s6     +1a40

#566rK   r   rc   s   @rJ   r9  r9  Z  s#    $
U\\ ell rK   r9  c                   8    e Zd ZU eed<   dZdZd ZdefdZ	d Z
y)	XmodPreTrainedModelrH   robertaTc                 l   t        |t        j                        rm|j                  j                  j                  d| j                  j                         |j                  %|j                  j                  j                          yyt        |t        j                        rz|j                  j                  j                  d| j                  j                         |j                  2|j                  j                  |j                     j                          yyt        |t        j                        rJ|j                  j                  j                          |j                  j                  j                  d       yt        |t              r%|j                  j                  j                          yy)zInitialize the weightsg        )meanstdNg      ?)r   r   rq   weightdatanormal_rH   initializer_rangebiaszero_r0   r#   r9   fill_
XmodLMHead)rG   modules     rJ   _init_weightsz!XmodPreTrainedModel._init_weightsp  s&   fbii( MM&&CT[[5R5R&S{{&  &&( '-MM&&CT[[5R5R&S!!-""6#5#56<<> .-KK""$MM$$S)
+KK""$ ,rK   r   c           	          || j                   j                  vr0t        |  d| dt        | j                   j                               || j                   _        y)z
        Set the default language code for the model. This is used when the language is not specified in the input.

        Args:
            language (`str`): The language code, such as `"en_XX"` or `"de_DE"`.
        z does not have an adapter for z. Supported languages: N)rH   r   rm   r   default_language)rG   r   s     rJ   set_default_languagez(XmodPreTrainedModel.set_default_language  s[     4;;000&6xj@WX\]a]h]h]r]rXsWtu  (0$rK   c                    t         j                  d       | j                  j                  j	                         D ]	  }d|_         t         j                  d       | j                  j                  j                  D ]x  }|j                  j                  0|j                  j                  j	                         D ]	  }d|_         |j                  j                  j	                         D ]	  }d|_         z y)z
        Freeze the embeddings and language adapters of the model. Usually, this is applied before the model is
        fine-tuned on a downstream task.
        zFreezing embeddingsFzFreezing adaptersN)r-  inforC  rY   
parametersrequires_gradencoderr  r   r   r   )rG   	parameterr  s      rJ   'freeze_embeddings_and_language_adaptersz;XmodPreTrainedModel.freeze_embeddings_and_language_adapters  s    
 	)*00;;= 	,I&+I#	,'(\\))// 	0E||..:!&!@!@!K!K!M 4I.3I+4"\\99DDF 0	*/	'0		0rK   N)r^   r_   r`   r   __annotations__base_model_prefixsupports_gradient_checkpointingrP  r   rS  rZ  r$  rK   rJ   rB  rB  i  s*    !&*#%$0S 00rK   rB  a0  
    The model can behave as an encoder (with only self-attention) as well as a decoder, in which case a layer of
    cross-attention is added between the self-attention layers, following the architecture described in *Attention is
    all you need*_ by Ashish Vaswani, Noam Shazeer, Niki Parmar, Jakob Uszkoreit, Llion Jones, Aidan N. Gomez, Lukasz
    Kaiser and Illia Polosukhin.

    To behave as an decoder the model needs to be initialized with the `is_decoder` argument of the configuration set
    to `True`. To be used in a Seq2Seq model, the model needs to initialized with both `is_decoder` argument and
    `add_cross_attention` set to `True`; an `encoder_hidden_states` is then expected as an input to the forward pass.

    .. _*Attention is all you need*: https://huggingface.co/papers/1706.03762
    )custom_introc            $           e Zd Zd fd	Zd Zd Zd Ze	 	 	 	 	 	 	 	 	 	 	 	 	 	 	 ddee	j                     dee	j                     dee	j                     dee	j                     d	ee	j                     d
ee	j                     dee	j                     dee	j                     dee	j                     deee	j                        dee   dee   dee   dee   dee	j                     deee	j                     ef   f d       Z xZS )	XmodModelc                     t         |   |       || _        t        |      | _        t        |      | _        |rt        |      nd| _        | j                          y)zv
        add_pooling_layer (bool, *optional*, defaults to `True`):
            Whether to add a pooling layer
        N)
r.   r/   rH   r!   rY   r  rX  r9  pooler	post_init)rG   rH   add_pooling_layerrI   s      rJ   r/   zXmodModel.__init__  sM    
 	 (0"6*,=j(4 	rK   c                 .    | j                   j                  S r   rY   r4   rG   s    rJ   get_input_embeddingszXmodModel.get_input_embeddings  s    ...rK   c                 &    || j                   _        y r   rf  )rG   rt   s     rJ   set_input_embeddingszXmodModel.set_input_embeddings  s    */'rK   c                     |j                         D ]7  \  }}| j                  j                  |   j                  j	                  |       9 y)z
        Prunes heads of the model. heads_to_prune: dict of {layer_num: list of heads to prune in this layer} See base
        class PreTrainedModel
        N)itemsrX  r  r
  r   )rG   heads_to_pruner  r   s       rJ   _prune_headszXmodModel._prune_heads  sE    
 +002 	CLE5LLu%//;;EB	CrK   rR   r   r{   r+   r(   r|   rS   r}   r  r  r   r   r!  r"  r   r   c                 ,   ||n| j                   j                  }||n| j                   j                  }||n| j                   j                  }| j                   j                  r||n| j                   j
                  }nd}||t        d      |#| j                  ||       |j                         }n!||j                         dd }nt        d      |\  }}||j                  n|j                  }d}|
5t        |
t              s|
d   d   j                  d   n|
j                         }|| j                   j                  t        d      t        | j                   j"                  d   j$                  j&                  j)                               }|j+                  | j                   j                        }|t-        j.                  ||	      z  }|t-        j.                  |||z   f|	      }|pt1        | j2                  d
      r4| j2                  j4                  ddd|f   }|j7                  ||      }|}n&t-        j8                  |t,        j:                  |      }| j=                  ||      }| j                   j                  rE|C|j                         \  }}}||f}|	t-        j.                  ||	      }	| j?                  |	      }nd}| jA                  || j                   jB                        }| j3                  |||||      }| j!                  |||||||
|||||      } | d   }!| jD                  | jE                  |!      nd}"|s
|!|"f| dd z   S tG        |!|"| jH                  | jJ                  | jL                  | jN                        S )  
        lang_ids (`torch.LongTensor` of shape `(batch_size, sequence_length)`, *optional*):
            Indices of the language adapters that should be activated for each sample, respectively. Default: the index
            that corresponds to `self.config.default_language`.
        NFzDYou cannot specify both input_ids and inputs_embeds at the same timer)   z5You have to specify either input_ids or inputs_embedsr   r   zPInput language unknown. Please call `XmodPreTrainedModel.set_default_language()`)rN   r+   rM   )rR   r(   r+   rS   rT   )r   r{   r|   r}   r  r  r   r   r!  r"  r   r   )r)  pooler_outputr  rz   r*  r+  )(rH   r   r!  use_return_dictrw   r   rm   %warn_if_padding_and_no_attention_maskrD   rN   r   r   r   get_seq_lengthrR  r   rX  r  r   r   r   r   r@   onesrQ   rY   r+   rB   rC   rE   get_extended_attention_maskinvert_attention_maskget_head_maskr  rb  r   r  rz   r*  r+  )#rG   rR   r   r{   r+   r(   r|   rS   r}   r  r  r   r   r!  r"  r   rU   r   rV   rN   rT   adapter_languagesdefault_lang_idrW   rX   extended_attention_maskencoder_batch_sizeencoder_sequence_lengthr   encoder_hidden_shapeencoder_extended_attention_maskembedding_outputencoder_outputssequence_outputr@  s#                                      rJ   rZ   zXmodModel.forward  s   0 2C1N-TXT_T_TqTq$8$D $++JjJj 	 &1%<k$++B]B];;!!%.%:	@U@UII ]%>cdd"66y.Q#..*K&',,.s3KTUU!,
J%.%:!!@T@T!"& "/59  "1%++B/$335 # {{++3 !stt $T\\%7%7%:%A%A%Q%Q%V%V%X Y/55dkk6R6RSO&Jv)NNH!"ZZ*jCY6Y)ZdjkN!t(89*.//*H*HKZK*X'3J3Q3QR\^h3i0!A!&[

SY!Z 150P0PQ_al0m ;;!!&;&G=R=W=W=Y: 7$68O#P %-).4HQW)X&.2.H.HI_.`+.2+ &&y$++2O2OP	??%)'#9 + 
 ,,2"7#B+/!5#) ' 
 *!,8<8OO4UY#]3oab6III;-'+;;)77&11,==
 	
rK   )T)NNNNNNNNNNNNNNN)r^   r_   r`   r/   rh  rj  rn  r   r   r@   r   
LongTensorr   r   r   r   r   r   rZ   rb   rc   s   @rJ   r`  r`    s    "/0C  -1/31515/3,0048<9==A$(,0/3&*15!A
ELL)A
 5++,A
 !.	A

 !.A
 u||,A
 ELL)A
  -A
  (5A
 !) 6A
 "$u'8'8"9:A
 D>A
 $D>A
 'tnA
 d^A
  !.!A
" 
uU\\"$PP	Q#A
 A
rK   r`  zQ
    X-MOD Model with a `language modeling` head on top for CLM fine-tuning.
    c            &       *    e Zd ZddgZ fdZd Zd Ze	 	 	 	 	 	 	 	 	 	 	 	 	 	 	 	 ddee	j                     dee	j                     dee	j                     d	ee	j                     d
ee	j                     dee	j                     dee	j                     dee	j                     dee	j                     dee	j                     deeee	j                           dee   dee   dee   dee   dee	j                     deee	j                     ef   f"d       Z xZS )XmodForCausalLMlm_head.decoder.weightlm_head.decoder.biasc                     t         |   |       |j                  st        j	                  d       t        |d      | _        t        |      | _        | j                          y )NzLIf you want to use `XmodLMHeadModel` as a standalone, add `is_decoder=True.`Frd  
r.   r/   rw   r-  warningr`  rC  rN  lm_headrc  rF   s     rJ   r/   zXmodForCausalLM.__init___  sL       NNij 5A!&) 	rK   c                 .    | j                   j                  S r   r  decoderrg  s    rJ   get_output_embeddingsz%XmodForCausalLM.get_output_embeddingsl      ||###rK   c                 &    || j                   _        y r   r  rG   new_embeddingss     rJ   set_output_embeddingsz%XmodForCausalLM.set_output_embeddingsp      -rK   rR   r   r{   r+   r(   r|   rS   r}   r  labelsr  r   r   r!  r"  r   r   c                    ||n| j                   j                  }|
d}| j                  |||||||||	||||||      }|d   }| j                  |      }d}|
* | j                  ||
fd| j                   j
                  i|}|s|f|dd z   }||f|z   S |S t        |||j                  |j                  |j                  |j                        S )aS  
        lang_ids (`torch.LongTensor` of shape `(batch_size, sequence_length)`, *optional*):
            Indices of the language adapters that should be activated for each sample, respectively. Default: the index
            that corresponds to `self.config.default_language`.
        labels (`torch.LongTensor` of shape `(batch_size, sequence_length)`, *optional*):
            Labels for computing the left-to-right language modeling loss (next word prediction). Indices should be in
            `[-100, 0, ..., config.vocab_size]` (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, XmodForCausalLM, AutoConfig
        >>> import torch

        >>> tokenizer = AutoTokenizer.from_pretrained("FacebookAI/xlm-roberta-base")
        >>> config = AutoConfig.from_pretrained("facebook/xmod-base")
        >>> config.is_decoder = True
        >>> model = XmodForCausalLM.from_pretrained("facebook/xmod-base", config=config)
        >>> model.set_default_language("en_XX")

        >>> inputs = tokenizer("Hello, my dog is cute", return_tensors="pt")
        >>> outputs = model(**inputs)

        >>> prediction_logits = outputs.logits
        ```NF)r   r{   r+   r(   r|   rS   r}   r  r  r   r   r!  r"  r   r   r1   rk   )losslogitsr  rz   r*  r+  )rH   rr  rC  r  loss_functionr1   r   r  rz   r*  r+  )rG   rR   r   r{   r+   r(   r|   rS   r}   r  r  r  r   r   r!  r"  r   kwargsr   r  prediction_scoreslm_lossr   s                          rJ   rZ   zXmodForCausalLM.forwards  s'   ^ &1%<k$++B]B]I,,))%'"7#9+/!5#)  
$ "!* LL9(d((!  ;;11 	G ')GABK7F,3,?WJ'KVK0$#33!//))$55
 	
rK   )NNNNNNNNNNNNNNNN)r^   r_   r`   _tied_weights_keysr/   r  r  r   r   r@   r  r   r   r   r   r   r   rZ   rb   rc   s   @rJ   r  r  V  s    34JK
$.  15/36:59371559=A>B-1EI$(,0/3&*15#[
E,,-[
 5++,[
 !!2!23	[

 !!1!12[
 u//0[
 E--.[
   1 12[
  ((9(9:[
 !)):): ;[
 ))*[
 "%e.?.?(@"AB[
 D>[
 $D>[
 'tn[
  d^![
" !.#[
& 
uU\\"$EE	F'[
 [
rK   r  c                        e Zd ZddgZ fdZd Zd Ze	 	 	 	 	 	 	 	 	 	 	 	 	 ddee	j                     dee	j                     dee	j                     d	ee	j                     d
ee	j                     dee	j                     dee	j                     dee	j                     dee	j                     dee	j                     dee   dee   dee   deee	j                     ef   fd       Z xZS )XmodForMaskedLMr  r  c                     t         |   |       |j                  rt        j	                  d       t        |d      | _        t        |      | _        | j                          y )NzkIf you want to use `XmodForMaskedLM` make sure `config.is_decoder=False` for bi-directional self-attention.Fr  r  rF   s     rJ   r/   zXmodForMaskedLM.__init__  sR     NN1
 !5A!&) 	rK   c                 .    | j                   j                  S r   r  rg  s    rJ   r  z%XmodForMaskedLM.get_output_embeddings  r  rK   c                 &    || j                   _        y r   r  r  s     rJ   r  z%XmodForMaskedLM.set_output_embeddings  r  rK   rR   r   r{   r+   r(   r|   rS   r}   r  r  r   r!  r"  r   c                    ||n| j                   j                  }| j                  |||||||||	|||      }|d   }| j                  |      }d}|
Ft	               } ||j                  d| j                   j                        |
j                  d            }|s|f|dd z   }||f|z   S |S t        |||j                  |j                        S )a  
        lang_ids (`torch.LongTensor` of shape `(batch_size, sequence_length)`, *optional*):
            Indices of the language adapters that should be activated for each sample, respectively. Default: the index
            that corresponds to `self.config.default_language`.
        labels (`torch.LongTensor` of shape `(batch_size, sequence_length)`, *optional*):
            Labels for computing the masked language modeling loss. Indices should be in `[-100, 0, ...,
            config.vocab_size]` (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]`
        N)r   r{   r+   r(   r|   rS   r}   r  r   r!  r"  r   r)   rk   r  r  rz   r*  )
rH   rr  rC  r  r   r   r1   r   rz   r*  )rG   rR   r   r{   r+   r(   r|   rS   r}   r  r  r   r!  r"  r   r  r  masked_lm_lossloss_fctr   s                       rJ   rZ   zXmodForMaskedLM.forward  s   4 &1%<k$++B]B],,))%'"7#9/!5#  
 "!* LL9')H%&7&<&<RAWAW&XZ`ZeZefhZijN')GABK7F3A3M^%.YSYY$!//))	
 	
rK   )NNNNNNNNNNNNN)r^   r_   r`   r  r/   r  r  r   r   r@   r  r   r   r   r   r   r   rZ   rb   rc   s   @rJ   r  r    sr   24JK $.  15/36:59371559=A>B-1,0/3&*:
E,,-:
 5++,:
 !!2!23	:

 !!1!12:
 u//0:
 E--.:
   1 12:
  ((9(9::
 !)):): ;:
 ))*:
 $D>:
 'tn:
 d^:
 
uU\\"N2	3:
 :
rK   r  c                   .     e Zd ZdZ fdZd Zd Z xZS )rN  z*Roberta Head for masked language modeling.c                    t         |           t        j                  |j                  |j                        | _        t        j                  |j                  |j                        | _        t        j                  |j                  |j                        | _
        t        j                  t        j                  |j                              | _        | j                  | j                  _        y r   )r.   r/   r   rq   r2   r   r9   r:   
layer_normr1   r  	Parameterr@   rC   rK  rF   s     rJ   r/   zXmodLMHead.__init__0  s    YYv1163E3EF
,,v'9'9v?T?TUyy!3!3V5F5FGLLV->->!?@	 IIrK   c                     | j                  |      }t        |      }| j                  |      }| j                  |      }|S r   )r   r   r  r  rG   featuresr  xs       rJ   rZ   zXmodLMHead.forward9  s;    JJx GOOA LLOrK   c                     | j                   j                  j                  j                  dk(  r| j                  | j                   _        y | j                   j                  | _        y )Nmeta)r  rK  rN   typerg  s    rJ   _tie_weightszXmodLMHead._tie_weightsC  sC     <<##((F2 $		DLL))DIrK   )r^   r_   r`   ra   r/   rZ   r  rb   rc   s   @rJ   rN  rN  -  s    4&*rK   rN  z
    X-MOD Model transformer with a sequence classification/regression head on top (a linear layer on top of the pooled
    output) e.g. for GLUE tasks.
    c                   ~    e Zd Z fdZe	 	 	 	 	 	 	 	 	 	 	 ddeej                     deej                     deej                     deej                     deej                     deej                     deej                     d	eej                     d
ee	   dee	   dee	   de
eej                     ef   fd       Z xZS )XmodForSequenceClassificationc                     t         |   |       |j                  | _        || _        t	        |d      | _        t        |      | _        | j                          y NFr  )	r.   r/   
num_labelsrH   r`  rC  XmodClassificationHead
classifierrc  rF   s     rJ   r/   z&XmodForSequenceClassification.__init__T  sJ      ++ 5A08 	rK   rR   r   r{   r+   r(   r|   rS   r  r   r!  r"  r   c                     ||n| j                   j                  }| j                  ||||||||	|
|
      }|d   }| j                  |      }d}|| j                   j                  | j
                  dk(  rd| j                   _        nl| j
                  dkD  rL|j                  t        j                  k(  s|j                  t        j                  k(  rd| j                   _        nd| j                   _        | j                   j                  dk(  rIt               }| j
                  dk(  r& ||j                         |j                               }n |||      }n| j                   j                  dk(  r=t               } ||j                  d| j
                        |j                  d            }n,| j                   j                  dk(  rt               } |||      }|s|f|d	d z   }||f|z   S |S t        |||j                   |j"                  
      S )a  
        lang_ids (`torch.LongTensor` of shape `(batch_size, sequence_length)`, *optional*):
            Indices of the language adapters that should be activated for each sample, respectively. Default: the index
            that corresponds to `self.config.default_language`.
        labels (`torch.LongTensor` of shape `(batch_size,)`, *optional*):
            Labels for computing the sequence classification/regression loss. Indices should be in `[0, ...,
            config.num_labels - 1]`. If `config.num_labels == 1` a regression loss is computed (Mean-Square loss), If
            `config.num_labels > 1` a classification loss is computed (Cross-Entropy).
        N	r   r{   r+   r(   r|   rS   r   r!  r"  r   r   
regressionsingle_label_classificationmulti_label_classificationr)   rk   r  )rH   rr  rC  r  problem_typer  r-   r@   rE   rn   r   squeezer   r   r   r   rz   r*  rG   rR   r   r{   r+   r(   r|   rS   r  r   r!  r"  r   r  r  r  r  r   s                     rJ   rZ   z%XmodForSequenceClassification.forward_  s   0 &1%<k$++B]B],,))%'/!5#  
 "!*1{{''/??a'/;DKK,__q(fllejj.HFLL\a\e\eLe/LDKK,/KDKK,{{''<7"9??a'#FNN$4fnn6FGD#FF3D))-JJ+-B @&++b/R))-II,./Y,F)-)9TGf$EvE'!//))	
 	
rK   NNNNNNNNNNN)r^   r_   r`   r/   r   r   r@   r  r   r   r   r   r   r   rZ   rb   rc   s   @rJ   r  r  L  s;   	  15/36:59371559-1,0/3&*H
E,,-H
 5++,H
 !!2!23	H

 !!1!12H
 u//0H
 E--.H
   1 12H
 ))*H
 $D>H
 'tnH
 d^H
 
uU\\"$<<	=H
 H
rK   r  c                   ~    e Zd Z fdZe	 	 	 	 	 	 	 	 	 	 	 ddeej                     deej                     deej                     deej                     deej                     deej                     deej                     d	eej                     d
ee	   dee	   dee	   de
eej                     ef   fd       Z xZS )XmodForMultipleChoicec                     t         |   |       t        |      | _        t	        j
                  |j                        | _        t	        j                  |j                  d      | _
        | j                          y )Nr   )r.   r/   r`  rC  r   r;   r<   r=   rq   r2   r  rc  rF   s     rJ   r/   zXmodForMultipleChoice.__init__  sV      (zz&"<"<=))F$6$6: 	rK   rR   r   r+   r{   r  r(   r|   rS   r   r!  r"  r   c                    ||n| j                   j                  }||j                  d   n|j                  d   }|!|j                  d|j	                  d            nd}|2|j                  |j	                  d      |j	                  d      z        nd}|!|j                  d|j	                  d            nd}|!|j                  d|j	                  d            nd}|!|j                  d|j	                  d            nd}|1|j                  d|j	                  d      |j	                  d            nd}| j                  ||||||||	|
|
      }|d   }| j                  |      }| j                  |      }|j                  d|      }d}|t               } |||      }|s|f|dd z   }||f|z   S |S t        |||j                  |j                        S )	a|  
        input_ids (`torch.LongTensor` of shape `(batch_size, num_choices, sequence_length)`):
            Indices of input sequence tokens in the vocabulary.

            Indices can be obtained using [`AutoTokenizer`]. See [`PreTrainedTokenizer.encode`] and
            [`PreTrainedTokenizer.__call__`] for details.

            [What are input IDs?](../glossary#input-ids)
        lang_ids (`torch.LongTensor` of shape `(batch_size, num_choices, sequence_length)`, *optional*):
            Indices of the language adapters that should be activated for each sample, respectively. Default: the index
            that corresponds to `self.config.default_language`.
        token_type_ids (`torch.LongTensor` of shape `(batch_size, num_choices, sequence_length)`, *optional*):
            Segment token indices to indicate first and second portions of the inputs. Indices are selected in `[0,
            1]`:

            - 0 corresponds to a *sentence A* token,
            - 1 corresponds to a *sentence B* token.

            [What are token type IDs?](../glossary#token-type-ids)
        labels (`torch.LongTensor` of shape `(batch_size,)`, *optional*):
            Labels for computing the multiple choice classification loss. Indices should be in `[0, ...,
            num_choices-1]` where `num_choices` is the size of the second dimension of the input tensors. (See
            `input_ids` above)
        position_ids (`torch.LongTensor` of shape `(batch_size, num_choices, sequence_length)`, *optional*):
            Indices of positions of each input sequence tokens in the position embeddings. Selected in the range `[0,
            config.max_position_embeddings - 1]`.

            [What are position IDs?](../glossary#position-ids)
        inputs_embeds (`torch.FloatTensor` of shape `(batch_size, num_choices, sequence_length, hidden_size)`, *optional*):
            Optionally, instead of passing `input_ids` you can choose to directly pass an embedded representation. This
            is useful if you want more control over how to convert `input_ids` indices into associated vectors than the
            model's internal embedding lookup matrix.
        Nr   r)   r   r   )	r   r(   r+   r{   r|   rS   r   r!  r"  rk   r  )rH   rr  r   r   rD   repeatrC  r=   r  r   r   rz   r*  )rG   rR   r   r+   r{   r  r(   r|   rS   r   r!  r"  num_choicesflat_input_idsflat_lang_idsflat_position_idsflat_token_type_idsflat_attention_maskflat_inputs_embedsr   r@  r  reshaped_logitsr  r  r   s                             rJ   rZ   zXmodForMultipleChoice.forward  s   ` &1%<k$++B]B],5,Aiooa(}GZGZ[\G]CLCXINN2,>?^bRZRf	q(9INN1<M(MNlpLXLdL--b,2C2CB2GHjnR`Rln11"n6I6I"6MNrvR`Rln11"n6I6I"6MNrv ( r=#5#5b#9=;M;Mb;QR 	 ,,"*..,/!5#  
  
]3/ ++b+6')HOV4D%''!"+5F)-)9TGf$EvE("!//))	
 	
rK   r  )r^   r_   r`   r/   r   r   r@   r  r   r   r   r   r   r   rZ   rb   rc   s   @rJ   r  r    s;     15/3596:-1371559,0/3&*]
E,,-]
 5++,]
 !!1!12	]

 !!2!23]
 ))*]
 u//0]
 E--.]
   1 12]
 $D>]
 'tn]
 d^]
 
uU\\"$==	>]
 ]
rK   r  c                   ~    e Zd Z fdZe	 	 	 	 	 	 	 	 	 	 	 ddeej                     deej                     deej                     deej                     deej                     deej                     deej                     d	eej                     d
ee	   dee	   dee	   de
eej                     ef   fd       Z xZS )XmodForTokenClassificationc                 d   t         |   |       |j                  | _        t        |d      | _        |j
                  |j
                  n|j                  }t        j                  |      | _	        t        j                  |j                  |j                        | _        | j                          y r  )r.   r/   r  r`  rC  classifier_dropoutr<   r   r;   r=   rq   r2   r  rc  rG   rH   r  rI   s      rJ   r/   z#XmodForTokenClassification.__init__  s      ++ 5A)/)B)B)NF%%TZTnTn 	 zz"45))F$6$68I8IJ 	rK   rR   r   r{   r+   r(   r|   rS   r  r   r!  r"  r   c                    ||n| j                   j                  }| j                  ||||||||	|
|
      }|d   }| j                  |      }| j	                  |      }d}|<t               } ||j                  d| j                        |j                  d            }|s|f|dd z   }||f|z   S |S t        |||j                  |j                        S )a  
        lang_ids (`torch.LongTensor` of shape `(batch_size, sequence_length)`, *optional*):
            Indices of the language adapters that should be activated for each sample, respectively. Default: the index
            that corresponds to `self.config.default_language`.
        labels (`torch.LongTensor` of shape `(batch_size, sequence_length)`, *optional*):
            Labels for computing the token classification loss. Indices should be in `[0, ..., config.num_labels - 1]`.
        Nr  r   r)   rk   r  )rH   rr  rC  r=   r  r   r   r  r   rz   r*  r  s                     rJ   rZ   z"XmodForTokenClassification.forward*  s    , &1%<k$++B]B],,))%'/!5#  
 "!*,,71')HFKKDOO<fkk"oNDY,F)-)9TGf$EvE$!//))	
 	
rK   r  )r^   r_   r`   r/   r   r   r@   r  r   r   r   r   r   r   rZ   rb   rc   s   @rJ   r  r    s-     15/36:59371559-1,0/3&*7
E,,-7
 5++,7
 !!2!23	7

 !!1!127
 u//07
 E--.7
   1 127
 ))*7
 $D>7
 'tn7
 d^7
 
uU\\"$99	:7
 7
rK   r  c                   (     e Zd ZdZ fdZd Z xZS )r  z-Head for sentence-level classification tasks.c                 Z   t         |           t        j                  |j                  |j                        | _        |j                  |j                  n|j                  }t        j                  |      | _	        t        j                  |j                  |j                        | _        y r   )r.   r/   r   rq   r2   r   r  r<   r;   r=   r  out_projr  s      rJ   r/   zXmodClassificationHead.__init__i  s    YYv1163E3EF
)/)B)B)NF%%TZTnTn 	 zz"45		&"4"4f6G6GHrK   c                     |d d dd d f   }| j                  |      }| j                  |      }t        j                  |      }| j                  |      }| j	                  |      }|S r>  )r=   r   r@   tanhr  r  s       rJ   rZ   zXmodClassificationHead.forwardr  sY    Q1WLLOJJqMJJqMLLOMM!rK   )r^   r_   r`   ra   r/   rZ   rb   rc   s   @rJ   r  r  f  s    7IrK   r  c                       e Zd Z fdZe	 	 	 	 	 	 	 	 	 	 	 	 ddeej                     deej                     deej                     deej                     deej                     deej                     deej                     d	eej                     d
eej                     dee	   dee	   dee	   de
eej                     ef   fd       Z xZS )XmodForQuestionAnsweringc                     t         |   |       |j                  | _        t        |d      | _        t        j                  |j                  |j                        | _        | j                          y r  )
r.   r/   r  r`  rC  r   rq   r2   
qa_outputsrc  rF   s     rJ   r/   z!XmodForQuestionAnswering.__init__  sU      ++ 5A))F$6$68I8IJ 	rK   rR   r   r{   r+   r(   r|   rS   start_positionsend_positionsr   r!  r"  r   c                 *   ||n| j                   j                  }| j                  ||||||||
||
      }|d   }| j                  |      }|j	                  dd      \  }}|j                  d      j                         }|j                  d      j                         }d}||	t        |j                               dkD  r|j                  d      }t        |	j                               dkD  r|	j                  d      }	|j                  d      }|j                  d|      }|	j                  d|      }	t        |      } |||      } |||	      }||z   dz  }|s||f|dd z   }||f|z   S |S t        ||||j                  |j                  	      S )
rp  Nr  r   r   r)   r   )ignore_indexrk   )r  start_logits
end_logitsrz   r*  )rH   rr  rC  r  r   r  r   r   rD   clampr   r   rz   r*  )rG   rR   r   r{   r+   r(   r|   rS   r  r  r   r!  r"  r   r  r  r  r  
total_lossignored_indexr  
start_lossend_lossr   s                           rJ   rZ   z XmodForQuestionAnswering.forward  s   * &1%<k$++B]B],,))%'/!5#  
 "!*1#)<<r<#: j#++B/::<''+668

&=+D?'')*Q."1"9"9""==%%'(1, - 5 5b 9(--a0M-33A}EO)//=AM']CH!,@J
M:H$x/14J"J/'!"+=F/9/EZMF*Q6Q+%!!//))
 	
rK   )NNNNNNNNNNNN)r^   r_   r`   r/   r   r   r@   r  r   r   r   r   r   r   rZ   rb   rc   s   @rJ   r  r  |  sT     15/36:593715596:48,0/3&*E
E,,-E
 5++,E
 !!2!23	E

 !!1!12E
 u//0E
 E--.E
   1 12E
 "%"2"23E
   0 01E
 $D>E
 'tnE
 d^E
 
uU\\"$@@	AE
 E
rK   r  c                     | j                  |      j                         }t        j                  |d      j	                  |      |z   |z  }|j                         |z   S )a  
    Replace non-padding symbols with their position numbers. Position numbers begin at padding_idx+1. Padding symbols
    are ignored. This is modified from fairseq's `utils.make_positions`.

    Args:
        x: torch.Tensor x:

    Returns: torch.Tensor
    r   r   )nern   r@   cumsumtype_asrE   )rR   r#   rT   maskincremental_indicess        rJ   rO   rO     sW     <<$((*D <<!4<<TBE[[_cc##%33rK   )r  r  r  r  r  r  r`  rB  )r   )Dra   r   typingr   r   r@   torch.utils.checkpointr   torch.nnr   r   r   activationsr
   r   cache_utilsr   r   
generationr   modeling_layersr   modeling_outputsr   r   r   r   r   r   r   r   modeling_utilsr   pytorch_utilsr   r   r   utilsr   r   configuration_xmodr   
get_loggerr^   r-  Moduler!   re   r   r   r   r   r   r  r  r9  rB  r`  r  r  rN  r  r  r  r  r  rO   __all__r$  rK   rJ   <module>r     sR     "    A A ' 5 ) 9	 	 	 . l l , * 
		H	%V=RYY V=t@.		 @.FRYY 5BII 5rryy ")) $/ /dI3* I3Xa
")) a
J  30/ 30 30l e
# e
e
P 
t
)? t

t
n V
) V
 V
t* *> V
$7 V
V
r j
/ j
 j
Z H
!4 H
 H
XRYY , R
2 R
 R
l4 	rK   