o
    	۷iQ                     @   s  d dl Z d dlmZmZmZ d dlZd dlmZ ddlm	Z	 ddl
mZmZ ddlmZ ddlmZ ddlmZmZmZ dd	lmZmZ dd
lmZmZ ddlmZmZ ddlmZ ddl m!Z!m"Z"m#Z# ddl$m%Z% ddl&m'Z' ddl(m)Z) G dd dej*Z+G dd dej*Z,G dd dej*Z-dej.de/dej.fddZ0	d;dej*dej.d ej.d!ej.d"eej. d#e1d$e1d%ee! fd&d'Z2d(d) Z3d<d*d+Z4G d,d- d-ej*Z5G d.d/ d/eZ6e"G d0d1 d1eZ7e"G d2d3 d3e7Z8e"G d4d5 d5e7eZ9G d6d7 d7ee7Z:G d8d9 d9ee7Z;g d:Z<dS )=    N)CallableOptionalUnion   )ACT2FN)CacheDynamicCache)GenerationMixin)create_causal_mask) 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   )HeliumConfigc                       s.   e Zd Zd fdd	Zdd Zdd Z  ZS )	HeliumRMSNormư>c                    s&   t    tt|| _|| _d S N)super__init__nn	Parametertorchonesweightvariance_epsilon)selfhidden_sizeeps	__class__ `/home/ubuntu/vllm_env/lib/python3.10/site-packages/transformers/models/helium/modeling_helium.pyr    0   s   

zHeliumRMSNorm.__init__c                 C   sR   |j }|tj}|djddd}|t|| j  }| jtj| |S )N   T)keepdim)	dtypetor#   float32powmeanrsqrtr&   r%   )r'   hidden_statesinput_dtypevariancer,   r,   r-   forward5   s
   zHeliumRMSNorm.forwardc                 C   s   t | jj d| j S )Nz, eps=)tupler%   shaper&   )r'   r,   r,   r-   
extra_repr<   s   zHeliumRMSNorm.extra_repr)r   )__name__
__module____qualname__r    r:   r=   __classcell__r,   r,   r*   r-   r   /   s    r   c                       sD   e Zd ZU ejed< ddef fddZe e	dd Z
  ZS )	HeliumRotaryEmbeddinginv_freqNconfigc                    s   t    t|drt|jtr|jd|jd| _nd| _|j| _	|j| _
|| _t| j | _| | j|\}| _| jd|dd | j| _d S )Nrope_scaling	rope_typetypedefaultrC   F)
persistent)r   r    hasattr
isinstancerE   dictgetrF   max_position_embeddingsmax_seq_len_cachedoriginal_max_seq_lenrD   r   rope_init_fnattention_scalingregister_bufferrC   original_inv_freq)r'   rD   devicerC   r*   r,   r-   r    C   s   
zHeliumRotaryEmbedding.__init__c           
      C   s   | j d d d d f  |jd dd|j}|d d d d d f  }t|jjtr6|jjdkr6|jjnd}t	j
|dd+ | |  dd}t	j||fdd	}| | j }| | j }	W d    n1 smw   Y  |j|jd
|	j|jd
fS )Nr   r/   r   mpscpuF)device_typeenabledr.   dim)r1   )rC   floatexpandr<   r2   rU   rK   rG   strr#   autocast	transposecatcosrR   sinr1   )
r'   xposition_idsinv_freq_expandedposition_ids_expandedrX   freqsembrb   rc   r,   r,   r-   r:   T   s   0&zHeliumRotaryEmbedding.forwardr   )r>   r?   r@   r#   Tensor__annotations__r   r    no_gradr   r:   rA   r,   r,   r*   r-   rB   @   s   
 
rB   c                       s$   e Zd Z fddZdd Z  ZS )	HeliumMLPc                    sx   t    || _|j| _|j| _tj| j| j|jd| _tj| j| j|jd| _	tj| j| j|jd| _
t|j | _d S )Nbias)r   r    rD   r(   intermediate_sizer!   Linearmlp_bias	gate_projup_proj	down_projr   
hidden_actact_fnr'   rD   r*   r,   r-   r    e   s   
zHeliumMLP.__init__c                 C   s$   |  | | || | }|S r   )ru   rw   rs   rt   )r'   rd   ru   r,   r,   r-   r:   o   s    zHeliumMLP.forward)r>   r?   r@   r    r:   rA   r,   r,   r*   r-   rm   d   s    
rm   r7   n_repreturnc                 C   s^   | j \}}}}|dkr| S | dddddddddf |||||} | ||| ||S )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)r<   r]   reshape)r7   ry   batchnum_key_value_headsslenhead_dimr,   r,   r-   	repeat_kvt   s
   0r           modulequerykeyvalueattention_maskscalingdropoutkwargsc                 K   s   t || j}t || j}	t||dd| }
|d ur3|d d d d d d d |jd f }|
| }
tjj|
dtj	d
|j}
tjj|
|| jd}
t|
|	}|dd }||
fS )Nr.   r   r/   )r[   r1   )ptrainingr   )r   num_key_value_groupsr#   matmulr`   r<   r!   
functionalsoftmaxr3   r2   r1   r   r   
contiguous)r   r   r   r   r   r   r   r   
key_statesvalue_statesattn_weightscausal_maskattn_outputr,   r,   r-   eager_attention_forward   s   
&r   c                 C   s>   | ddddf }| ddddf }t j| |fdddS )	z*Rotates half the hidden dims of the input..r   Nr.   r   r/   rZ   r   )r#   stackflatten)rd   x1x2r,   r,   r-   rotate_half   s   r   c                 C   s   | |}| |}|dd|jd d f jddd}|dd|jd d f jddd}| | t| |  }|| t||  }||fS )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.   rZ   )	unsqueezer<   repeat_interleaver   )qkrb   rc   re   unsqueeze_dimq_embedk_embedr,   r,   r-   apply_rotary_pos_emb   s   

$$r   c                       s   e Zd ZdZddedee f fddZeddd	d
		dde	j
dee	j
e	j
f dee	j
 dee dee	j dee dee	j
e	j
f fddZ  ZS )HeliumAttentionz=Multi-headed attention from 'Attention Is All You Need' paperNrD   	layer_idxc                    s   t    || _|| _t|d|j|j | _|j|j | _	dt
| j | _|j| _d| _tj|j|j| j |jd| _tj|j|j| j |jd| _tj|j|j| j |jd| _tj|j|jdd| _d S )Nr   r   Trn   F)r   r    rD   r   getattrr(   num_attention_headsr   r}   r   mathsqrtr   attention_dropout	is_causalr!   rq   attention_biasq_projk_projv_projo_projr'   rD   r   r*   r,   r-   r       s$   
zHeliumAttention.__init__past_key_valuepast_key_values4.58new_nameversionr7   position_embeddingsr   cache_positionr   rz   c                 K   s$  |j d d }g |d| jR }| ||dd}	| ||dd}
| ||dd}|\}}t|	|
||\}	}
|d urW|||d}||
|| j	|\}
}t
}| jjdkret| jj }|| |	|
||f| jsqdn| j| jd|\}}|jg |dR   }| |}||fS )Nr/   r   r.   )rc   rb   r   eagerr   )r   r   )r<   r   r   viewr`   r   r   r   updater   r   rD   _attn_implementationr   r   r   r   r{   r   r   )r'   r7   r   r   r   r   r   input_shapehidden_shapequery_statesr   r   rb   rc   cache_kwargsattention_interfacer   r   r,   r,   r-   r:      s8   


zHeliumAttention.forwardr   )NN)r>   r?   r@   __doc__r   r   intr    r   r#   rj   r;   r   
LongTensorr   r   r:   rA   r,   r,   r*   r-   r      s*    r   c                       s   e Zd Zddedee f fddZedddd					
		ddej	deej	 deej
 dee dee deej
 deeej	ej	f  dee dej	fddZ  ZS )HeliumDecoderLayerNrD   r   c                    sR   t    |j| _t||d| _t|| _t|j|jd| _	t|j|jd| _
d S )N)rD   r   r)   )r   r    r(   r   	self_attnrm   mlpr   rms_norm_epsinput_layernormpost_attention_layernormr   r*   r,   r-   r      s   

zHeliumDecoderLayer.__init__r   r   r   r   Fr7   r   re   	use_cacher   r   r   rz   c              
   K   s^   |}	|  |}| jd|||||||d|\}}
|	| }|}	| |}| |}|	| }|S )N)r7   r   re   r   r   r   r   r,   )r   r   r   r   )r'   r7   r   re   r   r   r   r   r   residual_r,   r,   r-   r:     s&   




zHeliumDecoderLayer.forwardr   )NNNFNN)r>   r?   r@   r   r   r   r    r   r#   rj   r   r   boolr;   r   r   r:   rA   r,   r,   r*   r-   r     s8    
	
r   c                   @   sH   e Zd ZU eed< dZdZdgZdgZdZ	dZ
dZdZdZeedZdS )HeliumPreTrainedModelrD   modelTr   r   )r7   
attentionsN)r>   r?   r@   r   rk   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_outputsr,   r,   r,   r-   r   5  s   
 
r   c                       s   e Zd Zdef fddZe e							ddeej	 deej
 deej	 dee d	eej d
eej	 dee dee defddZ  ZS )HeliumModelrD   c                    s   t     j| _ j| _t j j| j| _t	 fddt
 jD | _t j jd| _t | _d| _|   d S )Nc                    s   g | ]}t  |qS r,   )r   ).0r   rD   r,   r-   
<listcomp>Q  s    z(HeliumModel.__init__.<locals>.<listcomp>r   F)r   r    pad_token_idpadding_idx
vocab_sizer!   	Embeddingr(   embed_tokens
ModuleListrangenum_hidden_layerslayersr   r   normrB   
rotary_embgradient_checkpointing	post_initrx   r*   r   r-   r    J  s   
zHeliumModel.__init__N	input_idsr   re   r   inputs_embedsr   r   r   rz   c              	   K   s   |d u |d uA rt d|d u r| |}|r!|d u r!t| jd}|d u r=|d ur-| nd}	tj|	|	|jd  |jd}|d u rF|	d}t
| j|||||d}
|}| ||}| jd | jj D ]}||f|
||||d|}qb| |}t||dS )	Nz:You must specify exactly one of input_ids or inputs_embedsr   r   r   )rU   )rD   input_embedsr   r   r   re   )r   re   r   r   r   )last_hidden_stater   )
ValueErrorr   r   rD   get_seq_lengthr#   aranger<   rU   r   r
   r   r   r   r   r   )r'   r   r   re   r   r   r   r   r   past_seen_tokensr   r7   r   decoder_layerr,   r,   r-   r:   Z  sP   

	

zHeliumModel.forward)NNNNNNN)r>   r?   r@   r   r    r   r   r   r#   r   rj   r   FloatTensorr   r   r   r   r:   rA   r,   r,   r*   r-   r   H  s<    	
r   c                       s   e Zd ZdgZddiZddgdgfiZ fddZee										dd
e	e
j de	e
j de	e
j de	e de	e
j de	e
j de	e de	e
j deee
jf dee defddZ  ZS )HeliumForCausalLMzlm_head.weightlm_headcolwise_repr7   logitsc                    s@   t  | t|| _|j| _tj|j|jdd| _| 	  d S )NFrn   )
r   r    r   r   r   r!   rq   r(   r   r   rx   r*   r,   r-   r      s
   
zHeliumForCausalLM.__init__Nr   r   r   re   r   r   labelsr   r   logits_to_keepr   rz   c
              
   K   s   | j d|||||||d|
}|j}t|	trt|	 dn|	}| |dd|ddf }d}|durB| jd||| jjd|
}t	|||j
|j|jdS )a  
        Example:

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

        >>> model = HeliumForCausalLM.from_pretrained("google/helium-7b")
        >>> tokenizer = AutoTokenizer.from_pretrained("google/helium-7b")

        >>> prompt = "What is your favorite condiment?"
        >>> 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]
        "What is your favorite condiment?"
        ```)r   r   re   r   r   r   r   N)r   r   r   )lossr   r   r7   r   r,   )r   r   rK   r   slicer   loss_functionrD   r   r   r   r7   r   )r'   r   r   re   r   r   r   r   r   r   r   outputsr7   slice_indicesr   r   r,   r,   r-   r:     s0    zHeliumForCausalLM.forward)	NNNNNNNNr   )r>   r?   r@   _tied_weights_keys_tp_plan_pp_planr    r   r   r   r#   r   rj   r   r   r   r   r   r   r   r   r:   rA   r,   r,   r*   r-   r     sN    		
r   c                   @      e Zd ZdS )HeliumForSequenceClassificationNr>   r?   r@   r,   r,   r,   r-   r	        r	  c                   @   r  )HeliumForTokenClassificationNr
  r,   r,   r,   r-   r    r  r  )r   r   r   r	  r  )r   )Nr   )=r   typingr   r   r   r#   torch.nnr!   activationsr   cache_utilsr   r   
generationr	   masking_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_heliumr   Moduler   rB   rm   rj   r   r   r\   r   r   r   r   r   r   r   r   r	  r  __all__r,   r,   r,   r-   <module>   sh   $

!E.NK