
    bCi=?                        S r 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
  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  SSKJr  SSKJrJrJrJrJ r J!r!  SSK"J#r#  \RH                  " \%5      r& " S S\RN                  5      r( " S S\ 5      r)S r*S$S jr+ " S S\5      r, " S S\5      r- " S S\5      r. " S S \5      r/ " S! S"\5      r0/ S#Qr1g)%zPyTorch Cohere model.    )CallableOptionalUnionN)nn   )Cache)FlashAttentionKwargs)GradientCheckpointingLayer)BaseModelOutputWithPastCausalLMOutputWithPast)dynamic_rope_update)ALL_ATTENTION_FUNCTIONS)Unpack)TransformersKwargslogging)deprecate_kwarg   )LlamaAttentionLlamaForCausalLMLlamaMLP
LlamaModelLlamaRotaryEmbeddingeager_attention_forward   )CohereConfigc                   2   ^  \ rS rSrSU 4S jjrS rSrU =r$ )CohereLayerNorm5   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/james-whalen/.local/lib/python3.13/site-packages/transformers/models/cohere/modular_cohere.pyr!   CohereLayerNorm.__init__6   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      $ )NT)keepdimr   )	dtypetor#   float32meanpowrsqrtr&   r%   )r'   hidden_statesinput_dtyper5   variances        r,   forwardCohereLayerNorm.forward<   s    #))%((7!!"d!3!(--a055b$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   5   s    $- -r.   r   c                   L    \ rS rSr\R
                  " 5       \S 5       5       rSrg)CohereRotaryEmbeddingF   c                 0   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[        R                  " 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$ ! , (       d  f       N@= f)Nr   r0   r   mpscpuF)device_typeenabledr   dimr2   )inv_freqfloatexpandshape
isinstancedevicetypestrr#   autocast	transposerepeat_interleavecosattention_scalingsinr3   r2   )
r'   xposition_idsinv_freq_expandedposition_ids_expandedrJ   freqsembrZ   r\   s
             r,   r;   CohereRotaryEmbedding.forwardG   sB    !MM$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$cff177f&;;; DCs   BF
F N)	r=   r>   r?   r@   r#   no_gradr   r;   rA   rd   r.   r,   rE   rE   F   s"    
]]_<  <r.   rE   c                 |    U SS S S24   nU SSS S24   n[         R                  " U* U/SS9R                  S5      nU$ )N.r   r   r0   rL   )r#   stackflatten)r]   x1x2rot_xs       r,   rotate_halfrm   W   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$ )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.
rN   )r2   rP   	unsqueezerm   r3   )	qkrZ   r\   r^   unsqueeze_dimr2   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$;;;r.   c                   (   ^  \ rS rSrU 4S jrSrU =r$ )	CohereMLP}   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?   r@   r!   rA   rB   rC   s   @r,   rw   rw   }   s    Y Yr.   rw   c                   D  ^  \ 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$ )CohereAttention   z=Multi-headed attention from 'Attention Is All You Need' paperr   	layer_idxc                 &  > [         TU ]  X5        UR                  U l        U R                  (       a_  [        UR                  U R
                  4UR                  S9U l        [        UR                  U R
                  4UR                  S9U l	        g g )Nr(   r)   )
r    r!   use_qk_normr   num_attention_headshead_dimlayer_norm_epsq_normnum_key_value_headsk_normr'   r   r   r+   s      r,   r!   CohereAttention.__init__   sz    +!--)#77GVMbMbDK *#77GVMbMbDK r.   past_key_valuepast_key_values4.58new_nameversionr8   position_embeddingsattention_maskcache_positionkwargsreturnc                    UR                   S S n/ UQSPU R                  P7nU R                  U5      R                  U5      n	U R	                  U5      R                  U5      n
U R                  U5      R                  U5      nU R                  (       a"  U R                  U	5      n	U R                  U
5      n
U	R                  SS5      n	U
R                  SS5      n
UR                  SS5      nUu  p[        XX5      u  pUb$  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$                  U R&                  S.UD6u  nnUR(                  " / UQSP76 R+                  5       nU R-                  U5      nUU4$ )Nr0   r   r   )r\   rZ   r   eagerg        )dropoutscaling)rR   r   q_projviewk_projv_projr   r   r   rX   ru   updater   r   r   _attn_implementationr   trainingattention_dropoutr   reshape
contiguouso_proj)r'   r8   r   r   r   r   r   input_shapehidden_shapequery_states
key_statesvalue_statesrZ   r\   cache_kwargs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#[ &#&nUL'6'='=jX\XfXfht'u$J(?;;++w6"9$++:Z:Z"[$7	%
  $}}C$2H2HLL	%
 	%
!\ "));;;;FFHkk+.L((r.   )r   r   r   N)NN)r=   r>   r?   r@   __doc__r   r   intr!   r   r#   Tensortupler   
LongTensorr   r	   r;   rA   rB   rC   s   @r,   r   r      s    G
| 
 
 
 %0A6R ,0591)||1) #5<<#=>1) !.	1)
 "%1) !!1!121) -.1) 
u||Xell33	41) S1)r.   r   c                     ^  \ rS rSrS\S\4U 4S jjr\" SSSS9      SS	\R                  S
\
\R                     S\
\R                     S\
\   S\
\   S\
\R                     S\
\\R                  \R                  4      S\\   S\\R"                  \
\\R"                  \R"                  4      4   4S jj5       rSrU =r$ )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$.F<N<NU[UjUjkr.   r   r   r   r   r8   r   r^   	use_cacher   r   r   r   c                     Un	U R                  U5      nU R                  " SUUUUUUUS.UD6u  pU R                  U5      nX-   U-   nU$ )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`).
    cache_position (`torch.LongTensor` of shape `(sequence_length)`, *optional*):
        Indices depicting the position of the input sequence tokens in the sequence
    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.
)r8   r   r^   r   r   r   r   rd   )r   r   r   )r'   r8   r   r^   r   r   r   r   r   residualhidden_states_attention_hidden_states_mlps                r,   r;   CohereDecoderLayer.forward   st    > !,,];%)^^ 	&
')%+) 3	&
 	&
" !HH]3 :=NNr.   )r(   r   r   r   )NNNFNN)r=   r>   r?   r@   r   r   r!   r   r#   r   r   r   r   boolr   r   r	   FloatTensorr;   rA   rB   rC   s   @r,   r   r      s   l| l l %0A6R 2637+/$)59KO.||. !.. u//0	.
 "%. D>. !!1!12. &eELL%,,,F&GH. -.. 
u  (51B1BEDUDU1U+V"WW	X. S.r.   r   c                   0   ^  \ rS rSrS\4U 4S jjrSrU =r$ )CohereModeli  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S9U l	        [        UR                  UR                  S9U l        g s  snf )N)r   r   )r    r!   r   
ModuleListrangenum_hidden_layersr   layersrE   
rotary_embr   r(   r   normr   s      r,   r!   CohereModel.__init__  ss     mmDI&JbJbDcdDcy2Dcd
 0v>#1C1C&J_J_`	 es   B)r   r   r   )r=   r>   r?   r@   r   r!   rA   rB   rC   s   @r,   r   r     s    a| a ar.   r   c                   |  ^  \ rS rSrU 4S jr           SS\\R                     S\\R                     S\\R                     S\\	\
\\R                     4      S\\R                     S\\R                     S	\\   S
\\   S\\   S\\R                     S\	\\R                  4   S\\   S\4S jjrSrU =r$ )CohereForCausalLMi  c                    > [         TU ]  U5        [        U5      U l        UR                  U l        UR
                  U l        g r   )r    r!   r   modellogit_scaletie_word_embeddingsr   s     r,   r!   CohereForCausalLM.__init__  s8      (
!--#)#=#= r.   	input_idsr   r^   r   inputs_embedslabelsr   output_attentionsoutput_hidden_statesr   logits_to_keepr   r   c                    Ub  UOU R                   R                  nU	b  U	OU R                   R                  n	U R                  " SUUUUUUUU	U
S.	UD6nUR                  n[        U[        5      (       a  [        U* S5      OUnU R                  USS2USS24   5      nUU R                  -  nSnUb)  U R                  " SUX`R                   R                  S.UD6n[        UUUR                  UR                  UR                  S9$ )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, 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."
```N)	r   r   r^   r   r   r   r   r   r   )logitsr   
vocab_size)lossr   r   r8   
attentionsrd   )r   r   r   r   last_hidden_staterS   r   slicelm_headr   loss_functionr   r   r   r8   r   )r'   r   r   r^   r   r   r   r   r   r   r   r   r   outputsr8   slice_indicesr   r   s                     r,   r;   CohereForCausalLM.forward  s(   J 2C1N-TXT_T_TqTq$8$D $++JjJj 	
 ,0:: ,
)%+'/!5),
 ,
  118B>SV8W8W~ot4]kmA}a,?@A$***%%pVF{{OeOepiopD%#33!//))
 	
r.   )r   r   r   )NNNNNNNNNNr   )r=   r>   r?   r@   r!   r   r#   r   r   r   r   listr   r   r   r   r   r   r;   rA   rB   rC   s   @r,   r   r     s>   > 151537KO59-1$(,0/35934H
E,,-H
 !.H
 u//0	H

 "%tE4E4E/F(F"GHH
   1 12H
 ))*H
 D>H
 $D>H
 'tnH
 !!1!12H
 c5<</0H
 +,H
 
 H
 H
r.   r   )r   r   CoherePreTrainedModel)Nr   )2r   typingr   r   r   r#   r   cache_utilsr   modeling_flash_attention_utilsr	   modeling_layersr
   modeling_outputsr   r   modeling_rope_utilsr   modeling_utilsr   processing_utilsr   utilsr   r   utils.deprecationr   llama.modeling_llamar   r   r   r   r   r   configuration_coherer   
get_loggerr=   loggerModuler   rE   rm   ru   rw   r   r   r   r   __all__rd   r.   r,   <module>r     s   .  , ,     B 9 O 6 5 & 0 0  / 
		H	%-bii -"<0 <"<<Y YA)n A)H73 7ta* aO
( O
dr.   