
    cCiV                     \   S SK JrJrJr  S SKrS SKJr  SSKJr  SSK	J
r
Jr  SSKJr  SSKJr  SSKJr  SS	KJr  SS
KJrJrJr  SSKJrJr  SSKJrJr  SSKJrJ r   SSK!J"r"  SSK#J$r$J%r%J&r&  SSK'J(r(  SSK)J*r*  SSK+J,r,   " S S\RZ                  5      r. " S S\5      r/S\R`                  S\1S\R`                  4S jr2 S9S\RZ                  S\R`                  S\R`                  S\R`                  S \\R`                     S!\3S"\3S#\"\$   4S$ jjr4S% r5S:S& jr6 " S' S(\RZ                  5      r7\" S)5       " S* S+\RZ                  5      5       r8 " S, S-\RZ                  5      r9\% " S. S/\ 5      5       r:\% " S0 S1\:5      5       r;\% " S2 S3\:\5      5       r< " S4 S5\\:5      r= " S6 S7\\:5      r>/ S8Qr?g);    )CallableOptionalUnionN   )ACT2FN)CacheDynamicCache)GenerationMixin)use_kernel_forward_from_hub)create_causal_mask)FlashAttentionKwargs) GenericForSequenceClassificationGenericForTokenClassificationGradientCheckpointingLayer)BaseModelOutputWithPastCausalLMOutputWithPast)ROPE_INIT_FUNCTIONSdynamic_rope_update)ALL_ATTENTION_FUNCTIONSPreTrainedModel)Unpack)TransformersKwargsauto_docstringcan_return_tuple)deprecate_kwarg)check_model_inputs   )
Glm4Configc                   b   ^  \ rS rSrU 4S jrS\R                  S\R                  4S jrSrU =r	$ )Glm4MLP0   c                    > [         TU ]  5         Xl        [        R                  " UR
                  SUR                  -  SS9U l        [        R                  " UR                  UR
                  SS9U l        [        UR                     U l        g )N   Fbias)super__init__confignnLinearhidden_sizeintermediate_sizegate_up_proj	down_projr   
hidden_actactivation_fnselfr(   	__class__s     `/home/james-whalen/.local/lib/python3.13/site-packages/transformers/models/glm4/modeling_glm4.pyr'   Glm4MLP.__init__1   sn    IIf&8&8!f>V>V:V]bc6#;#;V=O=OV[\#F$5$56    hidden_statesreturnc                     U R                  U5      nUR                  SSS9u  p2X R                  U5      -  nU R                  U5      $ )Nr#   dim)r-   chunkr0   r.   )r2   r7   	up_statesgates       r4   forwardGlm4MLP.forward9   sH    %%m4	#//!/4 2 24 88	~~i((r6   )r0   r(   r.   r-   )
__name__
__module____qualname____firstlineno__r'   torchFloatTensorr@   __static_attributes____classcell__r3   s   @r4   r    r    0   s,    7)U%6%6 )5;L;L ) )r6   r    c                     ^  \ rS rSrS\S\4U 4S jjr\" SSSS9      SS	\R                  S
\
\R                     S\
\R                     S\
\   S\
\   S\
\R                     S\
\\R                  \R                  4      S\\   S\\R"                  \
\\R"                  \R"                  4      4   4S jj5       rSrU =r$ )Glm4DecoderLayerB   r(   	layer_idxc                   > [         TU ]  5         UR                  U l        [        XS9U l        [        U5      U l        [        UR                  UR                  S9U l	        [        UR                  UR                  S9U l
        [        UR                  UR                  S9U l        [        UR                  UR                  S9U l        g )N)r(   rN   eps)r&   r'   r+   Glm4Attention	self_attnr    mlpGlm4RMSNormrms_norm_epsinput_layernormpost_attention_layernormpost_self_attn_layernormpost_mlp_layernormr2   r(   rN   r3   s      r4   r'   Glm4DecoderLayer.__init__C   s    !--&fJ6?*6+=+=6CVCVW(3F4F4FFL_L_(`%(3F4F4FFL_L_(`%"-f.@.@fFYFY"Zr6   past_key_valuepast_key_values4.58new_nameversionr7   attention_maskposition_ids	use_cachecache_positionposition_embeddingskwargsr8   c                     Un	U R                  U5      nU R                  " SUUUUUUUS.UD6u  pU R                  U5      nX-   nUn	U R                  U5      nU R	                  U5      nU R                  U5      nX-   nU$ )N)r7   rc   rd   r^   re   rf   rg    )rW   rS   rY   rX   rT   rZ   )r2   r7   rc   rd   r^   re   rf   rg   rh   residual_s              r4   r@   Glm4DecoderLayer.forwardN   s     !,,];>> 	
')%+) 3	
 	
 55mD 0 55mD///> 0r6   )r+   rW   rT   rX   rZ   rY   rS   )NNNFNN)rB   rC   rD   rE   r   intr'   r   rF   Tensorr   
LongTensorr   booltupler   r   rG   r@   rH   rI   rJ   s   @r4   rL   rL   B   s   	[z 	[c 	[ %0A6R 2637+/$)59KO!||! !.! u//0	!
 "%! D>! !!1!12! &eELL%,,,F&GH! -.! 
u  (51B1BEDUDU1U+V"WW	X! S!r6   rL   r7   n_repr8   c                     U R                   u  p#pEUS:X  a  U $ U SS2SS2SSS2SS24   R                  X#XU5      n U R                  X#U-  XE5      $ )z
This is the equivalent of torch.repeat_interleave(x, dim=1, repeats=n_rep). The hidden states go from (batch,
num_key_value_heads, seqlen, head_dim) to (batch, num_attention_heads, seqlen, head_dim)
r   N)shapeexpandreshape)r7   rs   batchnum_key_value_headsslenhead_dims         r4   	repeat_kvr|   s   s_    
 2?1D1D.Ez!!Qa"23::5W\dlmM  e(CTTTr6   modulequerykeyvaluerc   scalingdropoutrh   c                 @   [        X R                  5      n[        X0R                  5      n	[        R                  " XR	                  SS5      5      U-  n
Ub"  US S 2S S 2S S 2S UR
                  S   24   nX-   n
[        R                  R                  U
S[        R                  S9R                  UR                  5      n
[        R                  R                  XU R                  S9n
[        R                  " X5      nUR	                  SS5      R                  5       nX4$ )Nr#   r   r:   )r<   dtype)ptrainingr   )r|   num_key_value_groupsrF   matmul	transposeru   r)   
functionalsoftmaxfloat32tor   r   r   
contiguous)r}   r~   r   r   rc   r   r   rh   
key_statesvalue_statesattn_weightscausal_maskattn_outputs                r4   eager_attention_forwardr      s     3 ; ;<JU$?$?@L<<';';Aq'ABWLL!$Q1.D
0@0@0D.D%DE#1==((2U]](SVVW\WbWbcL==((6??([L,,|:K''1-88:K$$r6   c                 x    U SSSS24   nU SSSS24   n[         R                  " U* U4SS9R                  S5      $ )	z*Rotates half the hidden dims of the input..r   Nr#   r   r:   r;   r   )rF   stackflatten)xx1x2s      r4   rotate_halfr      sJ    	
319B	
319B;;Ryb)11"55r6   c                    UR                  U5      nUR                  U5      nUSSUR                  S   S-  24   R                  SSS9nUSSUR                  S   S-  24   R                  SSS9nUR                  S   nU SSU24   U SUS24   pUSSU24   USUS24   pXr-  [        U5      U-  -   nX-  [        U	5      U-  -   n[        R
                  " X/SS9n[        R
                  " X/SS9nX4$ )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.
    position_ids (`torch.Tensor`, *optional*):
        Deprecated and unused.
    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.
.Nr:   r#   r;   )	unsqueezeru   repeat_interleaver   rF   cat)qkcossinrd   unsqueeze_dim
rotary_dimq_rotq_passk_rotk_passq_embedk_embeds                r4   apply_rotary_pos_embr      s6   ( --
&C
--
&C c'SYYr]a'''
(
:
:1"
:
EC
c'SYYr]a'''
(
:
:1"
:
EC 2Jc;J;&'3
+;)<6c;J;&'3
+;)<6 {{51C78G{{51C78G ii)r2Gii)r2Gr6   c                   >  ^  \ rS rSrSrSS\S\\   4U 4S jjjr\	" SSSS	9  SS
\
R                  S\\
R                  \
R                  4   S\\
R                     S\\   S\\
R                     S\\   S\\
R                  \
R                  4   4S jj5       rSrU =r$ )rR      z=Multi-headed attention from 'Attention Is All You Need' paperr(   rN   c                 <  > [         TU ]  5         Xl        X l        [	        USUR
                  UR                  -  5      U l        UR                  UR                  -  U l	        U R                  S-  U l
        UR                  U l        SU l        [        R                  " UR
                  UR                  U R                  -  UR                  S9U l        [        R                  " UR
                  UR                  U R                  -  UR                  S9U l        [        R                  " UR
                  UR                  U R                  -  UR                  S9U l        [        R                  " UR                  U R                  -  UR
                  SS9U l        g )Nr{   g      Tr$   F)r&   r'   r(   rN   getattrr+   num_attention_headsr{   ry   r   r   attention_dropout	is_causalr)   r*   attention_biasq_projk_projv_projo_projr[   s      r4   r'   Glm4Attention.__init__   s@   "
F4F4F&JdJd4de$*$>$>&B\B\$\!}}d*!'!9!9ii : :T]] JQWQfQf
 ii : :T]] JQWQfQf
 ii : :T]] JQWQfQf
 ii : :T]] JFL^L^ejkr6   r]   r^   r_   r`   r7   rg   rc   rf   rh   r8   c                 4   UR                   S S n/ UQSPU R                  P7nU R                  U5      R                  U5      R	                  SS5      n	U R                  U5      R                  U5      R	                  SS5      n
U R                  U5      R                  U5      R	                  SS5      nUu  p[        XX5      u  pUb$  XUS.nUR                  XU R                  U5      u  p[        nU R                  R                  S:w  a  [        U R                  R                     nU" U U	U
UU4U R                  (       d  SOU R                  U R                   S.UD6u  nnUR"                  " / UQSP76 R%                  5       nU R'                  U5      nUU4$ )Nr:   r   r#   )r   r   rf   eager        )r   r   )ru   r{   r   viewr   r   r   r   updaterN   r   r(   _attn_implementationr   r   r   r   rw   r   r   )r2   r7   rg   rc   r^   rf   rh   input_shapehidden_shapequery_statesr   r   r   r   cache_kwargsattention_interfacer   r   s                     r4   r@   Glm4Attention.forward   s    $))#2.88b8$--8{{=166|DNNqRST[[/44\BLLQPQR
{{=166|DNNqRST&#7RU#[ &#&nUL'6'='=jX\XfXfht'u$J(?;;++w6"9$++:Z:Z"[$7	%
  $}}C$2H2HLL	%
 	%
!\ "));;;;FFHkk+.L((r6   )r   r(   r{   r   r   rN   r   r   r   r   r   N)NN)rB   rC   rD   rE   __doc__r   r   rn   r'   r   rF   ro   rr   r   rp   r   r   r@   rH   rI   rJ   s   @r4   rR   rR      s    Glz lhsm l l* %0A6R ,059))||)) #5<<#=>)) !.	))
 "%)) !!1!12)) +,)) 
u||U\\)	*)) S))r6   rR   RMSNormc                   8   ^  \ rS rSrSU 4S jjrS rS rSrU =r$ )rU   i  c                    > [         TU ]  5         [        R                  " [        R
                  " U5      5      U l        X l        g)z*
Glm4RMSNorm is equivalent to T5LayerNorm
N)r&   r'   r)   	ParameterrF   onesweightvariance_epsilon)r2   r+   rQ   r3   s      r4   r'   Glm4RMSNorm.__init__  s/     	ll5::k#:; #r6   c                    UR                   nUR                  [        R                  5      nUR	                  S5      R                  SSS9nU[        R                  " X0R                  -   5      -  nU R                  UR                  U5      -  $ )Nr#   r:   T)keepdim)	r   r   rF   r   powmeanrsqrtr   r   )r2   r7   input_dtypevariances       r4   r@   Glm4RMSNorm.forward  sw    #))%((7 $$Q',,R,>%H?T?T4T(UU{{]--k:::r6   c                 ^    [        U R                  R                  5       SU R                   3$ )Nz, eps=)rr   r   ru   r   )r2   s    r4   
extra_reprGlm4RMSNorm.extra_repr   s*    ))*+6$2G2G1HIIr6   )r   r   )gư>)	rB   rC   rD   rE   r'   r@   r   rH   rI   rJ   s   @r4   rU   rU     s    $;J Jr6   rU   c                      ^  \ rS rSr% \R
                  \S'   SS\4U 4S jjjr\R                  " 5       \
S 5       5       rSrU =r$ )Glm4RotaryEmbeddingi$  inv_freqr(   c                   > [         TU ]  5         [        US5      (       aZ  [        UR                  [
        5      (       a;  UR                  R                  SUR                  R                  S5      5      U l        OSU l        UR                  U l	        UR                  U l
        Xl        [        U R                     U l        U R                  U R                  U5      u  o0l        U R                  SUSS9  U R                   U l        g )Nrope_scaling	rope_typetypedefaultr   F)
persistent)r&   r'   hasattr
isinstancer   dictgetr   max_position_embeddingsmax_seq_len_cachedoriginal_max_seq_lenr(   r   rope_init_fnattention_scalingregister_bufferr   original_inv_freq)r2   r(   devicer   r3   s       r4   r'   Glm4RotaryEmbedding.__init__'  s    6>**z&:M:Mt/T/T#0044[&BUBUBYBYZ`BabDN&DN"("@"@$*$B$B!/?+/+<+<T[[&+Q((ZeD!%r6   c                 b   U R                   S S S 2S 4   R                  5       R                  UR                  S   SS5      R	                  UR
                  5      nUS S 2S S S 24   R                  5       n[        UR
                  R                  [        5      (       a0  UR
                  R                  S:w  a  UR
                  R                  OSn[        R                  " USS9   UR                  5       UR                  5       -  R                  SS5      n[        R                  " Xf4SS	9nUR                  5       U R                  -  nUR                  5       U R                  -  n	S S S 5        WR	                  UR                   S
9W	R	                  UR                   S
94$ ! , (       d  f       N@= f)Nr   r:   r   mpscpuF)device_typeenabledr#   r;   )r   )r   floatrv   ru   r   r   r   r   strrF   autocastr   r   r   r   r   r   )
r2   r   rd   inv_freq_expandedposition_ids_expandedr   freqsembr   r   s
             r4   r@   Glm4RotaryEmbedding.forward8  sR    !MM$4-8>>@GGHZHZ[\H]_acdehhijiqiqr ,QaZ 8 > > @'1!((--'E'E!((--[`J`ahhmmfk^^UC&,,.1F1L1L1NNYYZ[]^_E))UN3C'')d444C'')d444C	 D vvAGGv$cff177f&;;; DCs   $BF  
F.)r   r(   r   r   r   r   r   r   )rB   rC   rD   rE   rF   ro   __annotations__r   r'   no_gradr   r@   rH   rI   rJ   s   @r4   r   r   $  s@    ll/z / /" ]]_<  <r6   r   c                   R    \ rS rSr% \\S'   SrSrS/rS/r	Sr
SrSrSrSr\\S.rSrg	)
Glm4PreTrainedModeliH  r(   modelTrL   r^   )r7   
attentionsrj   N)rB   rC   rD   rE   r   r   base_model_prefixsupports_gradient_checkpointing_no_split_modules_skip_keys_device_placement_supports_flash_attn_supports_sdpa_supports_flex_attn_can_compile_fullgraph_supports_attention_backendrL   rR   _can_record_outputsrH   rj   r6   r4   r  r  H  sQ    &*#+,#4"5N!"&)#r6   r  c                   "  ^  \ rS rSrS\4U 4S jjr\" 5       \       SS\\	R                     S\\	R                     S\\	R                     S\\   S\\	R                     S	\\	R                     S
\\   S\\   S\4S jj5       5       rSrU =r$ )	Glm4Modeli[  r(   c           	        > [         TU ]  U5        UR                  U l        UR                  U l        [
        R                  " UR                  UR                  U R                  5      U l        [
        R                  " [        UR                  5       Vs/ s H  n[        X5      PM     sn5      U l        [        UR                  UR                  S9U l        [#        US9U l        SU l        U R)                  5         g s  snf )NrP   r(   F)r&   r'   pad_token_idpadding_idx
vocab_sizer)   	Embeddingr+   embed_tokens
ModuleListrangenum_hidden_layersrL   layersrU   rV   normr   
rotary_embgradient_checkpointing	post_initr[   s      r4   r'   Glm4Model.__init__]  s     !.. ++LL):):F<N<NPTP`P`ammBGH`H`BabBaYf0Bab
   2 28K8KL	-V<&+# 	 cs   C?	input_idsrc   rd   r^   inputs_embedsrf   re   rh   r8   c           
      J   US L US L-  (       a  [        S5      eUc  U R                  U5      nU(       a  Uc  [        U R                  S9nUcD  Ub  UR	                  5       OSn	[
        R                  " XUR                  S   -   UR                  S9nUc  UR                  S5      n[        U R                  UUUUUS9n
UnU R                  X5      nU R                  S U R                  R                    H  nU" U4U
UUUUS.UD6nM     U R                  U5      n[        UUS9$ )	Nz:You must specify exactly one of input_ids or inputs_embedsr  r   r   )r   )r(   input_embedsrc   rf   r^   rd   )rc   rd   r^   rf   rg   )last_hidden_stater^   )
ValueErrorr  r	   r(   get_seq_lengthrF   arangeru   r   r   r   r  r  r  r  r   )r2   r!  rc   rd   r^   r"  rf   re   rh   past_seen_tokensr   r7   rg   decoder_layers                 r4   r@   Glm4Model.forwardm  sR    -t";<YZZ *.*;*;I*FM0*$++>O!CRC^==?de+0<< ]5H5H5K"KTaThTh,N )33A6L(;;&))+%
 &"oomJ![[)H4;;+H+HIM)*) /-$7 M J 		-0&++
 	
r6   )r  r  r  r  r  r  r  )NNNNNNN)rB   rC   rD   rE   r   r'   r   r   r   rF   rp   ro   r   rG   rq   r   r   r   r@   rH   rI   rJ   s   @r4   r  r  [  s    z    151537+/5959$(8
E,,-8
 !.8
 u//0	8

 "%8
   1 128
 !!1!128
 D>8
 +,8
 
!8
  8
r6   r  c                   |  ^  \ rS rSrS/rSS0rSS/S/40rU 4S jr\\	         SS\
\R                     S	\
\R                     S
\
\R                     S\
\   S\
\R                     S\
\R                     S\
\   S\
\R                     S\\\R                  4   S\\   S\\\4   4S jj5       5       rSrU =r$ )Glm4ForCausalLMi  zlm_head.weightlm_headcolwise_repr7   logitsc                    > [         TU ]  U5        [        U5      U l        UR                  U l        [
        R                  " UR                  UR                  SS9U l        U R                  5         g )NFr$   )
r&   r'   r  r  r  r)   r*   r+   r.  r  r1   s     r4   r'   Glm4ForCausalLM.__init__  sU     v&
 ++yy!3!3V5F5FUS 	r6   r!  rc   rd   r^   r"  labelsre   rf   logits_to_keeprh   r8   c
                 ~   U R                   " SUUUUUUUS.U
D6nUR                  n[        U	[        5      (       a  [	        U	* S5      OU	nU R                  USS2USS24   5      nSnUb)  U R                  " SXU R                  R                  S.U
D6n[        UUUR                  UR                  UR                  S9$ )a  
labels (`torch.LongTensor` of shape `(batch_size, sequence_length)`, *optional*):
    Labels for computing the masked language modeling loss. Indices should either be in `[0, ...,
    config.vocab_size]` or -100 (see `input_ids` docstring). Tokens with indices set to `-100` are ignored
    (masked), the loss is only computed for the tokens with labels in `[0, ..., config.vocab_size]`.

Example:

```python
>>> from transformers import AutoTokenizer, Glm4ForCausalLM

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

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

>>> # Generate
>>> generate_ids = model.generate(inputs.input_ids, max_length=30)
>>> tokenizer.batch_decode(generate_ids, skip_special_tokens=True, clean_up_tokenization_spaces=False)[0]
"Hey, are you conscious? Can you talk to me?\nI'm not conscious, but I can talk to you."
```)r!  rc   rd   r^   r"  re   rf   N)r0  r3  r  )lossr0  r^   r7   r  rj   )r  r%  r   rn   slicer.  loss_functionr(   r  r   r^   r7   r  )r2   r!  rc   rd   r^   r"  r3  re   rf   r4  rh   outputsr7   slice_indicesr0  r6  s                   r4   r@   Glm4ForCausalLM.forward  s    J ,0:: 	,
)%+')	,
 	,
  118B>SV8W8W~ot4]kmA}a,?@A%%pVt{{OeOepiopD%#33!//))
 	
r6   )r.  r  r  )	NNNNNNNNr   )rB   rC   rD   rE   _tied_weights_keys_tp_plan_pp_planr'   r   r   r   rF   rp   ro   r   rG   rq   r   rn   r   r   rr   r   r@   rH   rI   rJ   s   @r4   r-  r-    s;   *+=)H_-z:;H  151537+/59-1$(5934=
E,,-=
 !.=
 u//0	=

 "%=
   1 12=
 ))*=
 D>=
 !!1!12=
 c5<</0=
 +,=
 
u,,	-=
  =
r6   r-  c                       \ rS rSrSrg)Glm4ForSequenceClassificationi  rj   NrB   rC   rD   rE   rH   rj   r6   r4   r@  r@        r6   r@  c                       \ rS rSrSrg)Glm4ForTokenClassificationi  rj   NrA  rj   r6   r4   rD  rD    rB  r6   rD  )r  r  r-  r@  rD  )r   )Nr   )@typingr   r   r   rF   torch.nnr)   activationsr   cache_utilsr   r	   
generationr
   integrationsr   masking_utilsr   modeling_flash_attention_utilsr   modeling_layersr   r   r   modeling_outputsr   r   modeling_rope_utilsr   r   modeling_utilsr   r   processing_utilsr   utilsr   r   r   utils.deprecationr   utils.genericr   configuration_glm4r   Moduler    rL   ro   rn   r|   r   r   r   r   rR   rU   r   r  r  r-  r@  rD  __all__rj   r6   r4   <module>rX     s  , - ,   ! . ) 7 / B 
 P K F & I I 0 / *)bii )$.1 .b	UU\\ 	U# 	U%,, 	U& %II%<<% 
% <<	%
 U\\*% % % '(%46'TB)BII B)J Y'J")) J (J(!<")) !<H /  $ K
# K
 K
\ M
)? M
 M
`	$DFY 		!>@S 	r6   