+
    QV-j.-  ã                   ó"  € ^ RI Ht ^ RIt^ RIHt ^RIHt ^RIHtH	t	H
t
Ht ^RIHt ^RIHt ^RIHtHtHtHt ]P*                  ! ]4      t ! R R	]P0                  4      t] ! R
 R4      4       t] ! R R4      4       t] ! R R4      4       tR# )é    )ÚpartialN)ÚCache)ÚBaseModelOutputWithPastÚQuestionAnsweringModelOutputÚ SequenceClassifierOutputWithPastÚTokenClassifierOutput)Ú	AutoModel)ÚUnpack)ÚTransformersKwargsÚauto_docstringÚcan_return_tupleÚloggingc                   ó:   a a€ ] tR t^"t oRtRtV 3R ltRtVtV ;t	# )ÚGradientCheckpointingLayeraÜ  Base class for layers with gradient checkpointing.

This class enables gradient checkpointing functionality for a layer. By default, gradient checkpointing is disabled
(`gradient_checkpointing = False`). When `model.set_gradient_checkpointing()` is called, gradient checkpointing is
enabled by setting `gradient_checkpointing = True` and assigning a checkpointing function to `_gradient_checkpointing_func`.

Important:

    When using gradient checkpointing with `use_reentrant=True`, inputs that require gradients (e.g. hidden states)
    must be passed as positional arguments (`*args`) rather than keyword arguments to properly propagate gradients.

    Example:

        ```python
        >>> # Correct - hidden_states passed as positional arg
        >>> out = self.layer(hidden_states, attention_mask=attention_mask)

        >>> # Incorrect - hidden_states passed as keyword arg
        >>> out = self.layer(hidden_states=hidden_states, attention_mask=attention_mask)
        ```
Fc                óz  <€ V P                   '       Ed   V P                  '       Ed   R pV P                  P                  pRV R2pRV9   d    VR,          '       d   R VR&   VR,          pRpRV9   d   VR,          e   RVR&   VR,          pRpR	V9   d   VR	,          e   RVR	&   VR
,          pRpRV9   d   VR,          e   RVR&   VR,          pRpV'       d.   VP	                  R4      R,           p\
        P                  V4       V P                  ! \        \        SV `(  3/ VB .VO5!  # \        SV `(  ! V/ VB # )Fz7Caching is incompatible with gradient checkpointing in z	. SettingÚ	use_cachez `use_cache=False`,TÚpast_key_valueNz `past_key_value=None`,Úpast_key_valuesz `past_key_values=None`,Ú
layer_pastz `layer_past=None`,Ú,Ú.)Úgradient_checkpointingÚtrainingÚ	__class__Ú__name__ÚrstripÚloggerÚwarning_onceÚ_gradient_checkpointing_funcr   ÚsuperÚ__call__)ÚselfÚargsÚkwargsÚdo_warnÚ
layer_nameÚmessager   s   &*,   €Úm/Volumes/fast/ai/experiments/ui-tars-smoke/.venv/lib/python3.14/site-packages/transformers/modeling_layers.pyr!   Ú#GradientCheckpointingLayer.__call__;   sS  ø€ Ø×&×&Ñ&¨4¯=¯=©=ØˆGØŸ™×0Ñ0ˆJØOÐPZÈ|Ð[dÐeˆGà˜fÔ$¨°×)<Ô)<Ø&+��{Ñ#ØÐ0Õ0�Ø�ð   6Ô)¨fÐ5EÕ.FÒ.RØ+/�Ð'Ñ(ØÐ4Õ4�Ø�à  FÔ*¨vÐ6GÕ/HÒ/TØ,0�Ð(Ñ)ØÐ5Õ5�Ø�à˜vÔ%¨&°Õ*>Ò*JØ'+��|Ñ$ØÐ0Õ0�Ø�÷ Ø!Ÿ.™.¨Ó-°Õ3�Ü×#Ñ# GÔ,à×4Ò4´W¼U¹WÑ=MÑ5XÐQWÑ5XÐ`Ð[_Ó`Ð`Ü‰wÒ Ð0¨Ñ0Ð0ó    © )
r   Ú
__module__Ú__qualname__Ú__firstlineno__Ú__doc__r   r!   Ú__static_attributes__Ú__classdictcell__Ú__classcell__©r   Ú__classdict__s   @@r(   r   r   "   s   ù‡ € ñð, #Ð÷"1õ "1r*   r   c                   ó`   a a€ ] tR t^`t oRtV 3R lt]]RV3R lR ll4       4       tRt	Vt
V ;t# )Ú GenericForSequenceClassificationÚmodelc                ó<  <€ \         SV `  V4       VP                  V n        \        W P                  \
        P                  ! V4      4       \        P                  ! VP                  4       P                  V P                  R R7      V n        V P                  4        R# )F)ÚbiasN)r    Ú__init__Ú
num_labelsÚsetattrÚbase_model_prefixr	   Úfrom_configÚnnÚLinearÚget_text_configÚhidden_sizeÚscoreÚ	post_init©r"   Úconfigr   s   &&€r(   r:   Ú)GenericForSequenceClassification.__init__d   sl   ø€ Ü‰Ñ˜Ô Ø ×+Ñ+ˆŒä�×,Ñ,¬i×.CÒ.CÀFÓ.KÔLÜ—Y’Y˜v×5Ñ5Ó7×CÑCÀTÇ_Á_Ð[`ÔaˆŒ
ð 	�‰Ör*   c                ó&  <€ V ^8„  d   QhRS[ P                  R,          RS[ P                  R,          RS[ P                  R,          RS[R,          RS[ P                  R,          RS[ P                  R,          RS[R,          R	S[S[,          R
S[/	# ©é   Ú	input_idsNÚattention_maskÚposition_idsr   Úinputs_embedsÚlabelsr   r$   Úreturn)	ÚtorchÚ
LongTensorÚTensorr   ÚFloatTensorÚboolr
   r   r   )Úformatr4   s   "€r(   Ú__annotate__Ú-GenericForSequenceClassification.__annotate__p   s¨   ø€ ÷ 8
ñ 8
á×#Ñ# dÕ*ð8
ñ Ÿ™ tÕ+ð8
ñ ×&Ñ&¨Õ-ð	8
ñ
  �ð8
ñ ×(Ñ(¨4Õ/ð8
ñ × Ñ  4Õ'ð8
ñ ˜$•;ð8
ñ Ñ+Õ,ð8
ñ 
*ñ8
r*   c                óL  € \        W P                  4      ! V3R VRVRVRVRV/VB p	V	P                  p
V P                  V
4      pVe   VP                  ^ ,          pMVP                  ^ ,          pV P
                  P                  4       P                  f   V^8w  d   \        R4      hV P
                  P                  4       P                  f   RpMÚVe©   WP
                  P                  4       P                  8g  P                  VP                  \        P                  4      p\        P                  ! VP                  R,          VP                  \        P                  R7      pWþ,          P                  R4      pM.Rp\        P!                  V P"                  P$                   R24       V\        P                  ! WËP                  R	7      V3,          pRpVe   V P'                  W¶VV P
                  R
7      p\)        VVV	P*                  V	P,                  V	P.                  R7      # )rL   rM   r   rN   r   Nz=Cannot handle batch sizes > 1 if no padding token is defined.)ÚdeviceÚdtypezŠ will not detect padding tokens in `inputs_embeds`. Results may be unexpected if using padding tokens in conjunction with `inputs_embeds.`)rZ   )ÚlogitsrO   Úpooled_logitsrF   )Úlossr\   r   Úhidden_statesÚ
attentionséÿÿÿÿ)Úgetattrr=   Úlast_hidden_staterC   ÚshaperF   rA   Úpad_token_idÚ
ValueErrorÚtorZ   rQ   Úint32ÚarangeÚargmaxr   r   r   r   Úloss_functionr   r   r_   r`   )r"   rK   rL   rM   r   rN   rO   r   r$   Útransformer_outputsr_   r\   Ú
batch_sizeÚlast_non_pad_tokenÚnon_pad_maskÚtoken_indicesr]   r^   s   &&&&&&&&,         r(   ÚforwardÚ(GenericForSequenceClassification.forwardn   sþ  € ô 8?¸t×E[ÑE[Ô7\Øñ8
à)ð8
ð &ð8
ð ,ð	8
ð
 (ð8
ð  ð8
ð ñ8
Ðð ,×=Ñ=ˆØ—‘˜MÓ*ˆàÒ Ø"Ÿ™¨Õ+‰Jà&×,Ñ,¨QÕ/ˆJà�;‰;×&Ñ&Ó(×5Ñ5Ò=À*ÐPQÄ/ÜÐ\Ó]Ð]Ø�;‰;×&Ñ&Ó(×5Ñ5Ò=Ø!#ÑØÒ"à%¯©×)DÑ)DÓ)F×)SÑ)SÑS×WÑWÐX^×XeÑXeÔgl×grÑgrÓsˆLÜ!ŸLšL¨¯©¸Õ)<ÀVÇ]Á]ÔZ_×ZeÑZeÔfˆMØ"/Õ">×!FÑ!FÀrÓ!JÑà!#ÐÜ×ÑØ—>‘>×*Ñ*Ð+ð ,Zð Zôð
 œuŸ|š|¨J¿}¹}ÔMÐOaÐaÕbˆàˆØÒØ×%Ñ%¨VÐR_Ðhl×hsÑhsÐ%ÓtˆDä/ØØ Ø/×?Ñ?Ø-×;Ñ;Ø*×5Ñ5ô
ð 	
r*   )r;   rC   ©NNNNNNN©r   r,   r-   r.   r=   r:   r   r   rq   r0   r1   r2   r3   s   @@r(   r6   r6   `   s6   ù‡ € àÐõð Ø÷8
ñ 8
ó ó ÷8
ð 8
r*   r6   c                   ól   a a€ ] tR t^«t oRtV 3R ltR tR t]]	RV3R lR ll4       4       t
RtVtV ;t# )	ÚGenericForQuestionAnsweringr7   c                óæ   <€ \         SV `  V4       \        W P                  \        P
                  ! V4      4       \        P                  ! VP                  ^4      V n	        V P                  4        R# )rJ   N)r    r:   r<   r=   r	   r>   r?   r@   rB   Ú
qa_outputsrD   rE   s   &&€r(   r:   Ú$GenericForQuestionAnswering.__init__¯   sO   ø€ Ü‰Ñ˜Ô ä�×,Ñ,¬i×.CÒ.CÀFÓ.KÔLÜŸ)š) F×$6Ñ$6¸Ó:ˆŒð 	�‰Ör*   c                ó@   € \        W P                  4      P                  # ©N©rb   r=   Úembed_tokens)r"   s   &r(   Úget_input_embeddingsÚ0GenericForQuestionAnswering.get_input_embeddings¸   s   € Ü�t×3Ñ3Ó4×AÑAÐAr*   c                ó:   € V\        W P                  4      n        R # r{   r|   )r"   Úvalues   &&r(   Úset_input_embeddingsÚ0GenericForQuestionAnswering.set_input_embeddings»   s   € Ø=BŒ�×,Ñ,Ó-Ö:r*   c                ó:  <€ V ^8„  d   QhRS[ P                  R,          RS[ P                  R,          RS[ P                  R,          RS[R,          RS[ P                  R,          RS[ P                  R,          RS[ P                  R,          R	S[S[,          R
S[/	# )rJ   rK   NrL   rM   r   rN   Ústart_positionsÚend_positionsr$   rP   )rQ   rR   rS   r   rT   r
   r   r   )rV   r4   s   "€r(   rW   Ú(GenericForQuestionAnswering.__annotate__À   s¯   ø€ ÷ %
ñ %
á×#Ñ# dÕ*ð%
ñ Ÿ™ tÕ+ð%
ñ ×&Ñ&¨Õ-ð	%
ñ
  �ð%
ñ ×(Ñ(¨4Õ/ð%
ñ ×)Ñ)¨DÕ0ð%
ñ ×'Ñ'¨$Õ.ð%
ñ Ñ+Õ,ð%
ñ 
&ñ%
r*   c                ó®  € \        W P                  4      ! V3R VRVRVRV/VB p	V	P                  p
V P                  V
4      pVP	                  ^RR7      w  rÍVP                  R4      P                  4       pVP                  R4      P                  4       pRpVe   Ve   V P                  ! WÍWg3/ VB p\        VVVV	P                  V	P                  R7      # )rL   rM   r   rN   )ÚdimN)r^   Ústart_logitsÚ
end_logitsr_   r`   ra   )rb   r=   rc   rx   ÚsplitÚsqueezeÚ
contiguousrk   r   r_   r`   )r"   rK   rL   rM   r   rN   r…   r†   r$   ÚoutputsÚsequence_outputr\   rŠ   r‹   r^   s   &&&&&&&&,      r(   rq   Ú#GenericForQuestionAnswering.forward¾   sþ   € ô ,3°4×9OÑ9OÔ+PØñ,
à)ð,
ð &ð,
ð ,ð	,
ð
 (ð,
ð ñ,
ˆð "×3Ñ3ˆà—‘ Ó1ˆØ#)§<¡<°°r <Ó#:Ñ ˆØ#×+Ñ+¨BÓ/×:Ñ:Ó<ˆØ×'Ñ'¨Ó+×6Ñ6Ó8ˆ
àˆØÒ&¨=Ò+DØ×%Ò% lÀÑiÐbhÑiˆDä+ØØ%Ø!Ø!×/Ñ/Ø×)Ñ)ô
ð 	
r*   )rx   rs   )r   r,   r-   r.   r=   r:   r~   r‚   r   r   rq   r0   r1   r2   r3   s   @@r(   rv   rv   «   sB   ù‡ € àÐõòBòCð Ø÷%
ñ %
ó ó ÷%
ð %
r*   rv   c                   ó`   a a€ ] tR t^èt oRtV 3R lt]]RV3R lR ll4       4       tRt	Vt
V ;t# )ÚGenericForTokenClassificationr7   c                óâ  <€ \         SV `  V4       VP                  V n        \        W P                  \
        P                  ! V4      4       \        VR R4      e   VP                  pM\        VRR4      e   VP                  pMRp\        P                  ! V4      V n        \        P                  ! VP                  4       P                  VP                  4      V n        V P#                  4        R# )Úclassifier_dropoutNÚhidden_dropoutgš™™™™™¹?)r    r:   r;   r<   r=   r	   r>   rb   r•   r–   r?   ÚDropoutÚdropoutr@   rA   rB   rC   rD   )r"   rF   r•   r   s   && €r(   r:   Ú&GenericForTokenClassification.__init__ì   sº   ø€ Ü‰Ñ˜Ô Ø ×+Ñ+ˆŒä�×,Ñ,¬i×.CÒ.CÀFÓ.KÔLÜ�6Ð/°Ó6ÒBØ!'×!:Ñ!:ÑÜ�VÐ-¨tÓ4Ò@Ø!'×!6Ñ!6Ñà!$ÐÜ—z’zÐ"4Ó5ˆŒÜ—Y’Y˜v×5Ñ5Ó7×CÑCÀV×EVÑEVÓWˆŒ
ð 	�‰Ör*   c                ó&  <€ V ^8„  d   QhRS[ P                  R,          RS[ P                  R,          RS[ P                  R,          RS[R,          RS[ P                  R,          RS[ P                  R,          RS[R,          R	S[S[,          R
S[/	# rI   )	rQ   rR   rS   r   rT   rU   r
   r   r   )rV   r4   s   "€r(   rW   Ú*GenericForTokenClassification.__annotate__ÿ   s¨   ø€ ÷ !
ñ !
á×#Ñ# dÕ*ð!
ñ Ÿ™ tÕ+ð!
ñ ×&Ñ&¨Õ-ð	!
ñ
  �ð!
ñ ×(Ñ(¨4Õ/ð!
ñ × Ñ  4Õ'ð!
ñ ˜$•;ð!
ñ Ñ+Õ,ð!
ñ 
ñ!
r*   c                ó4  € \        W P                  4      ! V3R VRVRVRVRV/VB p	V	P                  p
V P                  V
4      p
V P	                  V
4      pRpVe   V P                  W¶V P                  4      p\        VVV	P                  V	P                  R7      # )rL   rM   r   rN   r   N)r^   r\   r_   r`   )
rb   r=   rc   r˜   rC   rk   rF   r   r_   r`   )r"   rK   rL   rM   r   rN   rO   r   r$   r�   r�   r\   r^   s   &&&&&&&&,    r(   rq   Ú%GenericForTokenClassification.forwardý   sÅ   € ô ,3°4×9OÑ9OÔ+PØñ,
à)ð,
ð &ð,
ð ,ð	,
ð
 (ð,
ð  ð,
ð ñ,
ˆð "×3Ñ3ˆØŸ,™, Ó7ˆØ—‘˜OÓ,ˆàˆØÒØ×%Ñ% f°d·k±kÓBˆDä$ØØØ!×/Ñ/Ø×)Ñ)ô	
ð 	
r*   )r˜   r;   rC   rs   rt   r3   s   @@r(   r“   r“   è   s6   ù‡ € àÐõð" Ø÷!
ñ !
ó ó ÷!
ð !
r*   r“   )Ú	functoolsr   rQ   Útorch.nnr?   Úcache_utilsr   Úmodeling_outputsr   r   r   r   Úmodels.autor	   Úprocessing_utilsr
   Úutilsr   r   r   r   Ú
get_loggerr   r   ÚModuler   r6   rv   r“   r+   r*   r(   Ú<module>r§      sž   ðõ ã Ý å ÷ó õ #Ý $ß PÓ Pð 
×	Ò	˜HÓ	%€ô;1 §¡ô ;1ð| ÷G
ð G
ó ðG
ðT ÷9
ð 9
ó ð9
ðx ÷7
ð 7
ó ò7
r*   