
    cCiV                     $   S SK JrJrJr  S SKrS SKJr  S SKJr  SSK	J
r
  SSKJrJr  SSKJr  SSKJr  SS	KJrJr  SS
KJr  SSKJrJr  SSKJrJr  SSKJrJr  SSK J!r!  SSK"J#r#J$r$  SSK%J&r&  SSK'J(r(  SSK)J*r*  \" S5       " S S\RV                  5      5       r,S\RZ                  S\.S\RZ                  4S jr/ S5S\RV                  S\RZ                  S\RZ                  S\RZ                  S\\RZ                     S \0S!\0S"\!\   4S# jjr1S6S$ jr2S% r3 " S& S'\RV                  5      r4 " S( S)\RV                  5      r5 " S* S+\5      r6 " S, S-\RV                  5      r7\# " S. S/\5      5       r8\# " S0 S1\85      5       r9\# " S2 S3\8\5      5       r:/ S4Qr;g)7    )CallableOptionalUnionN)TransformersKwargs   )ACT2FN)CacheDynamicCache)GenerationMixin)use_kernel_forward_from_hub)create_causal_mask!create_sliding_window_causal_mask)GradientCheckpointingLayer)BaseModelOutputWithPastCausalLMOutputWithPast)ROPE_INIT_FUNCTIONSdynamic_rope_update)ALL_ATTENTION_FUNCTIONSPreTrainedModel)Unpack)auto_docstringcan_return_tuple)deprecate_kwarg)check_model_inputs   )Olmo3ConfigRMSNormc                   8   ^  \ rS rSrSU 4S jjrS rS rSrU =r$ )Olmo3RMSNorm-   c                    > [         TU ]  5         [        R                  " [        R
                  " U5      5      U l        X l        g)z+
Olmo3RMSNorm is equivalent to T5LayerNorm
N)super__init__nn	Parametertorchonesweightvariance_epsilon)selfhidden_sizeeps	__class__s      b/home/james-whalen/.local/lib/python3.13/site-packages/transformers/models/olmo3/modeling_olmo3.pyr#   Olmo3RMSNorm.__init__/   s/     	ll5::k#:; #    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      $ )N   T)keepdim)	dtypetor&   float32powmeanrsqrtr)   r(   )r*   hidden_statesinput_dtypevariances       r.   forwardOlmo3RMSNorm.forward7   sw    #))%((7 $$Q',,R,>%H?T?T4T(UUm+//<<r0   c                 ^    [        U R                  R                  5       SU R                   3$ )Nz, eps=)tupler(   shaper)   )r*   s    r.   
extra_reprOlmo3RMSNorm.extra_repr>   s*    ))*+6$2G2G1HIIr0   )r)   r(   )gư>)	__name__
__module____qualname____firstlineno__r#   r>   rC   __static_attributes____classcell__r-   s   @r.   r   r   -   s    $=J Jr0   r   r;   n_repreturnc                     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)rB   expandreshape)r;   rL   batchnum_key_value_headsslenhead_dims         r.   	repeat_kvrU   B   s_    
 2?1D1D.Ez!!Qa"23::5W\dlmM  e(CTTTr0   modulequerykeyvalueattention_maskscalingdropoutkwargsc                 @   [        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$ )Nr2   r   r3   )dimr5   )ptrainingr   )rU   num_key_value_groupsr&   matmul	transposerB   r$   
functionalsoftmaxr7   r6   r5   r\   rb   
contiguous)rV   rW   rX   rY   rZ   r[   r\   r]   
key_statesvalue_statesattn_weightscausal_maskattn_outputs                r.   eager_attention_forwardrn   N   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$$r0   c                    U R                   UR                   pvUR                  U5      nUR                  U5      nX-  [        U 5      U-  -   nX-  [        U5      U-  -   n	UR                  U5      U	R                  U5      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.
)r5   	unsqueezerotate_halfr6   )
qkcossinposition_idsunsqueeze_dimq_typek_typeq_embedk_embeds
             r.   apply_rotary_pos_embr|   h   sv    ( WWaggF
--
&C
--
&Cw;q>C/0Gw;q>C/0G::fwzz&111r0   c                     U SSU R                   S   S-  24   nU SU R                   S   S-  S24   n[        R                  " U* U4SS9$ )z*Rotates half the hidden dims of the input..Nr3   r2   r`   )rB   r&   cat)xx1x2s      r.   rq   rq      sZ    	
3"!''"+"""	#B	
3q ""	#B99rc2YB''r0   c                   :  ^  \ rS rSrSrS\S\4U 4S jjr\" SSSS	9  SS
\	R                  S\\	R                  \	R                  4   S\\	R                     S\\   S\\	R                     S\\   S\\	R                  \\	R                     4   4S jj5       rSrU =r$ )Olmo3Attention   z=Multi-headed attention from 'Attention Is All You Need' paperconfig	layer_idxc                   > [         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
                  UR                  S9U l        [)        UR                  U R                  -  UR*                  5      U l        [)        UR                  U R                  -  UR*                  5      U l        UR0                  c   eUR0                  U   U l        U R2                  S:X  a  UR4                  U l        g S U l        g )NrT   g      Tbiassliding_attention)r"   r#   r   r   getattrr+   num_attention_headsrT   rR   rc   r[   attention_dropout	is_causalr$   Linearattention_biasq_projk_projv_projo_projr   rms_norm_epsq_normk_normlayer_typesattention_typesliding_windowr*   r   r   r-   s      r.   r#   Olmo3Attention.__init__   s   "
F4F4F&JdJd4de$*$>$>&B\B\$\!}}d*!'!9!9ii : :T]] JQWQfQf
 ii : :T]] JQWQfQf
 ii : :T]] JQWQfQf
 ii&&68J8JQWQfQf
 #6#=#=#MvObObc"6#=#=#MvObObc!!---$00;7;7J7JNa7af33gkr0   past_key_valuepast_key_values4.58new_nameversionr;   position_embeddingsrZ   cache_positionr]   rM   c                    UR                   S S n/ UQSPU R                  P7nU R                  U R                  U5      5      n	U R	                  U R                  U5      5      n
U R                  U5      nU	R                  U5      R                  SS5      n	U
R                  U5      R                  SS5      n
UR                  U5      R                  SS5      nUu  p[        XX5      u  pUb$  XUS.nUR                  XU R                  U5      u  p[        nU R                  R                  S:w  a  [        U R                  R                     nU" U U	U
UU4U R                   (       d  SOU R"                  U R$                  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$ )Nr3   r   r2   )ru   rt   r   eager        )r\   r[   r   )rB   rT   r   r   r   r   r   viewre   r|   updater   rn   r   _attn_implementationr   rb   r   r[   r   rP   rh   r   )r*   r;   r   rZ   r   r   r]   input_shapehidden_shapequery_statesri   rj   rt   ru   cache_kwargsattention_interfacerm   rk   s                     r.   r>   Olmo3Attention.forward   s    $))#2.88b8$--8{{4;;}#=>[[]!;<
{{=1#((6@@AF__\2<<QB
#((6@@AF&#7RU#[ &#&nUL'6'='=jX\XfXfht'u$J(?;;++w6"9$++:Z:Z"[$7
%
  $}}C$2H2HLL..
%
 
%
!\ "));;;;FFHkk+.L((r0   )r   r   r   rT   r   r   r   r   rc   r   r   r   r[   r   r   NN)rE   rF   rG   rH   __doc__r   intr#   r   r&   TensorrA   r   r	   
LongTensorr   r   r>   rI   rJ   rK   s   @r.   r   r      s    Gl{ ls l8 %0A6R ,059.)||.) #5<<#=>.) !.	.)
 "%.) !!1!12.) +,.) 
u||Xell33	4.) S.)r0   r   c                   .   ^  \ rS rSrU 4S jrS rSrU =r$ )Olmo3MLP   c                   > [         TU ]  5         Xl        UR                  U l        UR                  U l        [
        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	        [        UR                     U l        g NFr   )r"   r#   r   r+   intermediate_sizer$   r   	gate_projup_proj	down_projr   
hidden_actact_fnr*   r   r-   s     r.   r#   Olmo3MLP.__init__   s    !--!'!9!94#3#3T5K5KRWXyy!1!143I3IPUV4#9#94;K;KRWXV../r0   c                     U R                  U R                  U R                  U5      5      U R                  U5      -  5      nU$ )N)r   r   r   r   )r*   r   r   s      r.   r>   Olmo3MLP.forward   s6    NN4;;t~~a/@#ADLLQRO#ST	r0   )r   r   r   r   r+   r   r   )rE   rF   rG   rH   r#   r>   rI   rJ   rK   s   @r.   r   r      s    0 r0   r   c                   H  ^  \ 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                  4S jj5       rSrU =r$ )Olmo3DecoderLayer   r   r   c                   > [         TU ]  5         UR                  U l        [        XS9U l        [        U5      U l        [        UR                  UR                  S9U l	        [        UR                  UR                  S9U l
        g )N)r   r   r,   )r"   r#   r+   r   	self_attnr   mlpr   r   post_attention_layernormpost_feedforward_layernormr   s      r.   r#   Olmo3DecoderLayer.__init__   sj    !--'vKF#(4V5G5GVM`M`(a%*6v7I7IvObOb*c'r0   r   r   r   r   r;   rZ   rv   	use_cacher   r   r]   rM   c                     Un	U R                   " SUUUUUUUS.UD6u  pU R                  U5      nX-   nUn	U R                  U5      nU R                  U5      nX-   nU$ )N)r;   rZ   rv   r   r   r   r    )r   r   r   r   )r*   r;   rZ   rv   r   r   r   r   r]   residual_s              r.   r>   Olmo3DecoderLayer.forward   s     !>> 	
')%+) 3	
 	
 55mD 0 !/77F 0r0   )r+   r   r   r   r   )NNNFNN)rE   rF   rG   rH   r   r   r#   r   r&   r   r   r   r	   boolrA   r   r   r>   rI   rJ   rK   s   @r.   r   r      s    d{ ds d %0A6R 2637+/$)59KO|| !. u//0	
 "% D> !!1!12 &eELL%,,,F&GH +, 
 Sr0   r   c                      ^  \ rS rSr% \R
                  \S'   SS\S\\	   4U 4S jjjr
\R                  " 5       \S 5       5       rSrU =r$ )	Olmo3RotaryEmbeddingi  inv_freqr   	rope_typec                 B  > [         TU ]  5         Ub  X0l        Or[        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                  c   eUR                  U l	        UR                  U l
        Xl        [        U R                     U l        U R                  U R                  U5      u  o@l        U R                  SUSS9  U R                   U l        g )Nrope_scalingr   typedefaultr   F)
persistent)r"   r#   r   hasattr
isinstancer   dictgetmax_position_embeddingsmax_seq_len_cachedoriginal_max_seq_lenr   r   rope_init_fnattention_scalingregister_bufferr   original_inv_freq)r*   r   devicer   r   r-   s        r.   r#   Olmo3RotaryEmbedding.__init__  s     &NV^,,F<O<OQU1V1V#0044[&BUBUBYBYZ`BabDN&DN~~)))"("@"@$*$B$B!/?+/+<+<T[[&+Q((ZeD!%r0   c                    U R                   S S S 2S 4   R                  5       R                  UR                  S   SS5      R	                  UR
                  5      nUS S 2S S S 24   R                  5       n[        UR
                  R                  [        5      (       a0  UR
                  R                  S:w  a  UR
                  R                  OSn[        R                  " USS9   UR                  5       UR                  5       -  R                  SS5      n[        R                  " Xf4SS	9nUR                  5       U R                  -  nUR                  5       U R                  -  n	X4sS S S 5        $ ! , (       d  f       g = f)
Nr   r3   r   mpscpuF)device_typeenabledr2   r~   )r   floatrO   rB   r6   r   r   r   strr&   autocastre   r   rt   r   ru   )
r*   r   rv   inv_freq_expandedposition_ids_expandedr   freqsembrt   ru   s
             r.   r>   Olmo3RotaryEmbedding.forward0  s)    !MM$4-8>>@GGHZHZ[\H]_acdehhijiqiqr ,QaZ 8 > > @'1!((--'E'E!((--[`J`ahhmmfk^^UC&,,.1F1L1L1NNYYZ[]^_E))UN3C'')d444C'')d444C8 DCCs   $BE22
F )r   r   r   r   r   r   r   r   )rE   rF   rG   rH   r&   r   __annotations__r   r   r   r#   no_gradr   r>   rI   rJ   rK   s   @r.   r   r     sK    ll/{ /HSM / /* ]]_
  
r0   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	)
Olmo3PreTrainedModeli?  r   modelTr   r   )r;   
attentionsr   N)rE   rF   rG   rH   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_backendr   r   _can_record_outputsrI   r   r0   r.   r   r   ?  sQ    &*#,-#4"5N!"&*$r0   r   c                   "  ^  \ rS rSrS\4U 4S jjr\" 5       \       SS\\	R                     S\\	R                     S\\	R                     S\\   S\\	R                     S	\\	R                     S
\\   S\\   S\4S jj5       5       rSrU =r$ )
Olmo3ModeliR  r   c           	      F  > [         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        SU l        [
        R$                  " ['        USS9['        US9S.5      U l        U R+                  5         g s  snf )Nr   Fr   )r   r   r   r   full_attention)r"   r#   pad_token_idpadding_idx
vocab_sizer$   	Embeddingr+   embed_tokens
ModuleListrangenum_hidden_layersr   layersr   r   normgradient_checkpointing
ModuleDictr   rotary_embs	post_initr   s      r.   r#   Olmo3Model.__init__T  s     !.. ++LL):):F<N<NPTP`P`ammCHIaIaCbcCbiv1Cbc
 !!3!39L9LM	&+#==%9S\%]"6f"E
 	 ds   D	input_idsrZ   rv   r   inputs_embedsr   r   r]   rM   c           
      (   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=n
[        5      (       d*  U R                  UUUUUS.n[        S0 UD6[        S0 UD6S.n
UnU R                  S   " X5      U R                  S	   " X5      S
.nU R                  S U R                  R                     H>  nU" U4XR"                  R$                     UUUXR"                  R$                     S.UD6nM@     U R'                  U5      n[)        UUS9$ )Nz:You must specify exactly one of input_ids or inputs_embedsr  r   r   )r   )r   input_embedsrZ   r   r   rv   )r  r   r   r  r  )rZ   rv   r   r   r   )last_hidden_stater   r   )
ValueErrorr  r
   r   get_seq_lengthr&   arangerB   r   rp   r   r   r   r   r  r  r  r   r   r  r   )r*   r  rZ   rv   r   r   r   r   r]   past_seen_tokenscausal_mask_mappingmask_kwargsr;   position_embeddings_mappingdecoder_layers                  r.   r>   Olmo3Model.forwardi  s    -t";<YZZ *.*;*;I*FM0*$++>O!CRC^==?de+0<< ]5H5H5K"KTaThTh,N )33A6L ?-FF ++ -"0"0#2 ,K #5"C{"C%F%U%U#
 &!%!1!12E!F}!c"../?@]'
#
 "[[)H4;;+H+HIM)23J3J3Y3YZ) /-$?@W@W@f@f$g M J 		-0&++
 	
r0   )r  r  r  r  r  r  r  )NNNNNNN)rE   rF   rG   rH   r   r#   r   r   r   r&   r   r   r	   FloatTensorr   r   r   r   r>   rI   rJ   rK   s   @r.   r  r  R  s    { *  151537+/5959$(C
E,,-C
 !.C
 u//0	C

 "%C
   1 12C
 !!1!12C
 D>C
 +,C
 
!C
  C
r0   r  c                   r  ^  \ 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S jj5       5       rSrU =r$ )Olmo3ForCausalLMi  zlm_head.weightlm_headcolwise_repr;   logitsc                    > [         TU ]  U5        [        U5      U l        UR                  U l        [
        R                  " UR                  UR                  SS9U l        U R                  5         g r   )
r"   r#   r  r   r  r$   r   r+   r0  r  r   s     r.   r#   Olmo3ForCausalLM.__init__  sU     '
 ++yy!3!3V5F5FUS 	r0   r  rZ   rv   r   r   labelsr   r   logits_to_keepr]   rM   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$ )ao  
Example:

```python
>>> from transformers import AutoTokenizer, Olmo3ForCausalLM

>>> model = Olmo3ForCausalLM.from_pretrained("meta-olmo3/Olmo3-2-7b-hf")
>>> tokenizer = AutoTokenizer.from_pretrained("meta-olmo3/Olmo3-2-7b-hf")

>>> 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  rZ   rv   r   r   r   r   N)r2  r5  r  )lossr2  r   r;   r   r   )r   r#  r   r   slicer0  loss_functionr   r  r   r   r;   r   )r*   r  rZ   rv   r   r   r5  r   r   r6  r]   outputsr;   slice_indicesr2  r8  s                   r.   r>   Olmo3ForCausalLM.forward  s    @ ,0:: 	,
)%+')	,
 	,
  118B>SV8W8W~ot4]kmA}a,?@A%%pVt{{OeOepiopD%#33!//))
 	
r0   )r0  r   r  )	NNNNNNNNr   )rE   rF   rG   rH   _tied_weights_keys_tp_plan_pp_planr#   r   r   r   r&   r   r   r	   r-  r   r   r   r   r   r   r>   rI   rJ   rK   s   @r.   r/  r/    s0   *+=)H_-z:;H  151537+/59-1$(59348
E,,-8
 !.8
 u//0	8

 "%8
   1 128
 ))*8
 D>8
 !!1!128
 c5<</08
 +,8
 
 8
  8
r0   r/  )r/  r  r   )r   )Nr   )<typingr   r   r   r&   torch.nnr$   transformers.utils.genericr   activationsr   cache_utilsr	   r
   
generationr   integrationsr   masking_utilsr   r   modeling_layersr   modeling_outputsr   r   modeling_rope_utilsr   r   modeling_utilsr   r   processing_utilsr   utilsr   r   utils.deprecationr   utils.genericr   configuration_olmo3r   Moduler   r   r   rU   r   rn   r|   rq   r   r   r   r   r   r  r/  __all__r   r0   r.   <module>rT     s  , - ,   9 ! . ) 7 R 9 O K F & 5 0 / , Y'J299 J (J(	UU\\ 	U# 	U%,, 	U& %II%<<% 
% <<	%
 U\\*% % % '(%428(N)RYY N)bryy  )2 )X$299 $N ?  $ [
% [
 [
| H
+_ H
 H
V Er0   