
    3j6                        S r SSKJr  SSK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  SS
KJr  SSKJr  SSKJr  SSKJrJrJrJr  SSKJrJr  SSKJrJrJrJ r J!r!J"r"  SSK#J$r$  \RJ                  " \&5      r' " S S\RP                  5      r) " S S\!5      r*S r+S$S jr, " S S\5      r-\ " S S\5      5       r. " S S\
5      r/ " S S \ 5      r0 " S! S"\5      r1/ S#Qr2g)%zPyTorch Cohere model.    )CallableN)nn   )Cache)FlashAttentionKwargs)GradientCheckpointingLayer)BaseModelOutputWithPastCausalLMOutputWithPast)dynamic_rope_update)ALL_ATTENTION_FUNCTIONS)Unpack)TransformersKwargsauto_docstringcan_return_tuplelogging)maybe_autocastno_inherit_decorator   )LlamaAttentionLlamaForCausalLMLlamaMLP
LlamaModelLlamaRotaryEmbeddingeager_attention_forward   )CohereConfigc                   2   ^  \ rS rSrSU 4S jjrS rSrU =r$ )CohereLayerNorm4   c                    > [         TU ]  5         [        R                  " [        R
                  " U5      5      U l        X l        g)zcThe hidden size can be a tuple or an int. The tuple is used for QKNorm to normalize across head_dimN)super__init__r   	Parametertorchonesweightvariance_epsilon)selfhidden_sizeepsbias	__class__s       c/home/wildlama/miniconda3/lib/python3.13/site-packages/transformers/models/cohere/modular_cohere.pyr"   CohereLayerNorm.__init__5   s-    ll5::k#:; #    c                    UR                   nUR                  [        R                  5      nUR	                  SSS9nX-
  R                  S5      R	                  SSS9nX-
  [        R                  " X@R                  -   5      -  nU R                  R                  [        R                  5      U-  nUR                  U5      $ )NT)keepdimr   )	dtypetor$   float32meanpowrsqrtr'   r&   )r(   hidden_statesinput_dtyper6   variances        r-   forwardCohereLayerNorm.forward;   s    #))%((7!!"d!3!(--a055b$5G&-XH]H]=]1^^u}}5E,,r/   )r'   r&   )Ngh㈵>F)__name__
__module____qualname____firstlineno__r"   r<   __static_attributes____classcell__r,   s   @r-   r   r   4   s    $- -r/   r   c                   L    \ rS rSr\R
                  " 5       \S 5       5       rSrg)CohereRotaryEmbeddingE   c                    U R                   S S S 2S 4   R                  5       R                  UR                  S   SS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[        USS9   UR                  5       UR                  5       -  R                  SS5      n[        R                  " US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$ ! , (       d  f       N@= f)Nr   r1   r   mpscpuF)device_typeenabledr   dimr3   )inv_freqfloatexpandshape
isinstancedevicetypestrr   	transposer$   repeat_interleavecosattention_scalingsinr4   r3   )
r(   xposition_idsinv_freq_expandedposition_ids_expandedrK   freqsembrZ   r\   s
             r-   r<   CohereRotaryEmbedding.forwardF   s>    !MM$4-8>>@GGHZHZ[\H]_acde ,QaZ 8 > > @'1!((--'E'E!((--[`J`ahhmmfkUC&,,.1F1L1L1NNYYZ[]^_E))%;C'')d444C'')d444C	 D vvAGGv$cff177f&;;; DCs    BE<<
F
 N)	r>   r?   r@   rA   r$   no_gradr   r<   rB   rd   r/   r-   rF   rF   E   s"    
]]_<  <r/   rF   c                 |    U SS S S24   nU SSS S24   n[         R                  " U* U/SS9R                  S5      nU$ )N.r   r   r1   rM   )r$   stackflatten)r]   x1x2rot_xs       r-   rotate_halfrm   V   sL    	
3!8B	
319BKK"b	r*2226ELr/   c                 &   U R                   nU R                  5       n UR                  5       nUR                  U5      nUR                  U5      nX-  [        U 5      U-  -   nX-  [        U5      U-  -   nUR	                  US9UR	                  US94$ )aI  Applies Rotary Position Embedding to the query and key tensors.

Args:
    q (`torch.Tensor`): The query tensor.
    k (`torch.Tensor`): The key tensor.
    cos (`torch.Tensor`): The cosine part of the rotary embedding.
    sin (`torch.Tensor`): The sine part of the rotary embedding.
    unsqueeze_dim (`int`, *optional*, defaults to 1):
        The 'unsqueeze_dim' argument specifies the dimension along which to unsqueeze cos[position_ids] and
        sin[position_ids] so that they can be properly broadcasted to the dimensions of q and k. For example, note
        that cos[position_ids] and sin[position_ids] have the shape [batch_size, seq_len, head_dim]. Then, if q and
        k have the shape [batch_size, heads, seq_len, head_dim], then setting unsqueeze_dim=1 makes
        cos[position_ids] and sin[position_ids] broadcastable to the shapes of q and k. Similarly, if q and k have
        the shape [batch_size, seq_len, heads, head_dim], then set unsqueeze_dim=2.
Returns:
    `tuple(torch.Tensor)` comprising of the query and key tensors rotated using the Rotary Position Embedding.
rO   )r3   rQ   	unsqueezerm   r4   )qkrZ   r\   unsqueeze_dimr3   q_embedk_embeds           r-   apply_rotary_pos_embru   ^   s    $ GGE		A		A
--
&C
--
&Cw;q>C/0Gw;q>C/0G::E:"GJJUJ$;;;r/   c                   (   ^  \ rS rSrU 4S jrSrU =r$ )	CohereMLPz   c                 >  > [         TU ]  U5        [        R                  " U R                  U R
                  SS9U l        [        R                  " U R                  U R
                  SS9U l        [        R                  " U R
                  U R                  SS9U l        g )NF)r+   )	r!   r"   r   Linearr)   intermediate_size	gate_projup_proj	down_projr(   configr,   s     r-   r"   CohereMLP.__init__{   ss     4#3#3T5K5KRWXyy!1!143I3IPUV4#9#94;K;KRWXr/   )r~   r|   r}   )r>   r?   r@   rA   r"   rB   rC   rD   s   @r-   rw   rw   z   s    Y Yr/   rw   c                     ^  \ rS rSrSrSS\S\S-  4U 4S jjjr SS\R                  S\
\R                  \R                  4   S	\R                  S-  S
\S-  S\\   S\
\R                  \R                  S-  4   4S jjrSrU =r$ )CohereAttention   z=Multi-headed attention from 'Attention Is All You Need' paperNr   	layer_idxc                 &  > [         TU ]  X5        UR                  U l        U R                  (       a_  [        UR                  U R
                  4UR                  S9U l        [        UR                  U R
                  4UR                  S9U l	        g g Nr)   r*   )
r!   r"   use_qk_normr   num_attention_headshead_dimlayer_norm_epsq_normnum_key_value_headsk_normr(   r   r   r,   s      r-   r"   CohereAttention.__init__   sz    +!--)#77GVMbMbDK *#77GVMbMbDK r/   r9   position_embeddingsattention_maskpast_key_valueskwargsreturnc                 z   UR                   S S n/ UQSPU R                  P7nU R                  U5      R                  U5      nU R	                  U5      R                  U5      n	U R                  U5      R                  U5      n
U R                  (       a"  U R                  U5      nU R                  U	5      n	UR                  SS5      nU	R                  SS5      n	U
R                  SS5      n
Uu  p[        XX5      u  pUb  UR                  XU R                  5      u  p[        R                  " U R                  R                   ["        5      nU" U UU	U
U4U R$                  (       d  SOU R&                  U R(                  S.UD6u  pUR*                  " / UQSP76 R-                  5       nU R/                  U5      nX4$ )Nr1   r   r   g        )dropoutscaling)rS   r   q_projviewk_projv_projr   r   r   rX   ru   updater   r   get_interfacer   _attn_implementationr   trainingattention_dropoutr   reshape
contiguouso_proj)r(   r9   r   r   r   r   input_shapehidden_shapequery_states
key_statesvalue_statesrZ   r\   attention_interfaceattn_outputattn_weightss                   r-   r<   CohereAttention.forward   s    $))#2.88b8$--8{{=166|D[[/44\B
{{=166|D;;|4LZ0J#--a3))!Q/
#--a3&#7RU#[ &'6'='=jX\XfXf'g$J(?(M(MKK,,.E)
 %8	%
  $}}C$2H2HLL	%
 	%
! "));;;;FFHkk+.((r/   )r   r   r   N)r>   r?   r@   rA   __doc__r   intr"   r$   Tensortupler   r   r   r<   rB   rC   rD   s   @r-   r   r      s    G
| 
d
 
 
" )-.)||.) #5<<#=>.) t+	.)
 .) -..) 
u||U\\D00	1.) .)r/   r   c                   T  ^  \ rS rSrS\S\4U 4S jjr     SS\R                  S\R                  S-  S\R                  S-  S	\
S-  S
\S-  S\\R                  \R                  4   S-  S\\   S\\R                  \\R                  \R                  4   S-  4   4S jjrSrU =r$ )CohereDecoderLayer   r   r   c                    > [         TU ]  5         UR                  U l        [        XS9U l        [        U5      U l        [        UR                  UR                  S9U l	        g )N)r   r   r   )
r!   r"   r)   r   	self_attnrw   mlpr   r   input_layernormr   s      r-   r"   CohereDecoderLayer.__init__   sP    !--(LV$.F<N<NU[UjUjkr/   Nr9   r   r^   r   	use_cacher   r   r   c           
          UnU R                  U5      nU R                  " SUUUUUUS.UD6u  pU R                  U5      nX-   U-   nU$ )a<  
Args:
    hidden_states (`torch.FloatTensor`): input to the layer of shape `(batch, seq_len, embed_dim)`
    attention_mask (`torch.FloatTensor`, *optional*):
        attention mask of size `(batch_size, sequence_length)` if flash attention is used or `(batch_size, 1,
        query_sequence_length, key_sequence_length)` if default attention is used.
    past_key_values (`Cache`, *optional*): cached past key and value projection states
    output_attentions (`bool`, *optional*):
        Whether or not to return the attentions tensors of all attention layers. See `attentions` under
        returned tensors for more detail.
    use_cache (`bool`, *optional*):
        If set to `True`, `past_key_values` key value states are returned and can be used to speed up decoding
        (see `past_key_values`).
    position_embeddings (`tuple[torch.FloatTensor, torch.FloatTensor]`, *optional*):
        Tuple containing the cosine and sine positional embeddings of shape `(batch_size, seq_len, head_dim)`,
        with `head_dim` being the embedding dimension of each attention head.
)r9   r   r^   r   r   r   rd   )r   r   r   )r(   r9   r   r^   r   r   r   r   residualhidden_states_attention_hidden_states_mlps               r-   r<   CohereDecoderLayer.forward   sq    6 !,,];%)^^ &
')%+ 3&
 &
" !HH]3 :=NNr/   )r)   r   r   r   )NNNFN)r>   r?   r@   rA   r   r   r"   r$   r   
LongTensorr   boolr   r   r   FloatTensorr<   rB   rC   rD   s   @r-   r   r      s    l| l l /304(,!&HL*||* t+* &&-	*
 * $;* #5<<#=>E* -.* 
u  %(9(95;L;L(L"MPT"TT	U* *r/   r   c                   0   ^  \ rS rSrS\4U 4S jjrSrU =r$ )CohereModel   r   c           	        > [         TU ]  U5        [        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        g s  snf r   )r!   r"   r   
ModuleListrangenum_hidden_layersr   layersr   r)   r   normr   s      r-   r"   CohereModel.__init__   sg     mmDI&JbJbDcdDcy2Dcd
 $1C1C&J_J_`	 es   A=)r   r   )r>   r?   r@   rA   r   r"   rB   rC   rD   s   @r-   r   r      s    a| a ar/   r   c                   0  ^  \ rS rSrU 4S jr\\        SS\R                  S-  S\R                  S-  S\R                  S-  S\
S-  S\R                  S-  S	\R                  S-  S
\S-  S\\R                  -  S\\   S\4S jj5       5       rSrU =r$ )CohereForCausalLMi  c                    > [         TU ]  U5        [        U5      U l        UR                  U l        UR
                  U l        g r   )r!   r"   r   modellogit_scaletie_word_embeddingsr   s     r-   r"   CohereForCausalLM.__init__  s8      (
!--#)#=#= r/   N	input_idsr   r^   r   inputs_embedslabelsr   logits_to_keepr   r   c	           
         U R                   " S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XR                  -  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$ )au  
Example:

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

>> model = CohereForCausalLM.from_pretrained("CohereForAI/c4ai-command-r-v01")
>> tokenizer = AutoTokenizer.from_pretrained("CohereForAI/c4ai-command-r-v01")

>> 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   r   r^   r   r   r   N)logitsr   
vocab_size)lossr   r   r9   
attentionsrd   )r   last_hidden_staterT   r   slicelm_headr   loss_functionr   r   r
   r   r9   r   )r(   r   r   r^   r   r   r   r   r   r   outputsr9   slice_indicesr   r   s                  r-   r<   CohereForCausalLM.forward  s    > ,0:: ,
)%+',
 ,
  118B>SV8W8W~ot4]kmA}a,?@A***%%pVt{{OeOepiopD%#33!//))
 	
r/   )r   r   r   )NNNNNNNr   )r>   r?   r@   rA   r"   r   r   r$   r   r   r   r   r   r   r   r   r
   r<   rB   rC   rD   s   @r-   r   r     s    >  .2.204(,26*.!%-.6
##d*6
 t+6
 &&-	6

 6
 ((4/6
   4'6
 $;6
 ell*6
 +,6
 
 6
  6
r/   r   )r   r   CoherePreTrainedModel)r   )3r   collections.abcr   r$   r   cache_utilsr   modeling_flash_attention_utilsr   modeling_layersr   modeling_outputsr	   r
   modeling_rope_utilsr   modeling_utilsr   processing_utilsr   utilsr   r   r   r   utils.genericr   r   llama.modeling_llamar   r   r   r   r   r   configuration_coherer   
get_loggerr>   loggerModuler   rF   rm   ru   rw   r   r   r   r   __all__rd   r/   r-   <module>r      s   ,  $     B 9 O 6 5 & R R A  / 
		H	%-bii -"<0 <"<8Y Y =)n =) =)@23 2ja* a?
( ?
Dr/   