o
    iM                     @   s  d dl mZmZmZ d dlZd dlmZ d dlm  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 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$m%Z% ddl&m'Z' ddl(m)Z) G dd dej*Z+G dd dej*Z,dd Z-dej.de/dej.fddZ0	d7dej*dej.d ej.d!ej.d"eej. d#e1d$e1d%ee! fd&d'Z2d8d(d)Z3G d*d+ d+ej*Z4G d,d- d-eZ5G d.d/ d/ej*Z6e"G d0d1 d1eZ7e"G d2d3 d3e7Z8e"G d4d5 d5e7eZ9g d6Z:dS )9    )CallableOptionalUnionN   )ACT2FN)CacheDynamicCache)GenerationMixin)create_causal_mask)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   )
OlmoConfigc                       s@   e Zd ZdZdeddf fddZdejdejfdd	Z  Z	S )
OlmoLayerNormz/LayerNorm but with no learnable weight or bias.hidden_sizereturnNc                    s   t    |f| _d S N)super__init__normalized_shape)selfr   	__class__ c/home/ubuntu/veenaModal/venv/lib/python3.10/site-packages/transformers/models/olmo/modeling_olmo.pyr      s   
zOlmoLayerNorm.__init__hidden_statesc                 C   s,   |j }tj|jtjd| jd d dd|S )N)dtypegh㈵>)eps)r'   F
layer_normtotorchfloat32r    )r!   r&   
orig_dtyper$   r$   r%   forward#   s    zOlmoLayerNorm.forward)
__name__
__module____qualname____doc__intr   r,   Tensorr/   __classcell__r$   r$   r"   r%   r      s    r   c                       s$   e Zd Z fddZdd Z  ZS )OlmoMLPc                    sr   t    || _|j| _|j| _tj| j| jdd| _tj| j| jdd| _tj| j| jdd| _	t
|j | _d S NFbias)r   r   configr   intermediate_sizennLinear	gate_projup_proj	down_projr   
hidden_actact_fnr!   r;   r"   r$   r%   r   +   s   
zOlmoMLP.__init__c                 C   s$   |  | | || | }|S r   )rA   rC   r?   r@   )r!   xrA   r$   r$   r%   r/   5   s    zOlmoMLP.forward)r0   r1   r2   r   r/   r6   r$   r$   r"   r%   r7   *   s    
r7   c                 C   sH   | dd| j d d f }| d| j d d df }tj| |fddS )z*Rotates half the hidden dims of the input..N   dim)shaper,   cat)rE   x1x2r$   r$   r%   rotate_half:   s   rN   r&   n_repr   c                 C   s^   | j \}}}}|dkr| S | dddddddddf |||||} | ||| ||S )z
    This is the equivalent of torch.repeat_interleave(x, dim=1, repeats=n_rep). The hidden states go from (batch,
    num_key_value_heads, seqlen, head_dim) to (batch, num_attention_heads, seqlen, head_dim)
    r   N)rJ   expandreshape)r&   rO   batchnum_key_value_headsslenhead_dimr$   r$   r%   	repeat_kvA   s
   0rV           modulequerykeyvalueattention_maskscalingdropoutkwargsc                 K   s   t || j}t || j}	t||dd| }
|d ur3|d d d d d d d |jd f }|
| }
tjj|
dtj	d
|j}
tjj|
|| jd}
t|
|	}|dd }||
fS )NrG   r   rF   )rI   r'   )ptrainingr   )rV   num_key_value_groupsr,   matmul	transposerJ   r=   
functionalsoftmaxr-   r+   r'   r^   rb   
contiguous)rX   rY   rZ   r[   r\   r]   r^   r_   
key_statesvalue_statesattn_weightscausal_maskattn_outputr$   r$   r%   eager_attention_forwardM   s   
&rn   c           
      C   s^   | j |j }}||}||}| | 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.
    )r'   	unsqueezerN   r+   )
qkcossinposition_idsunsqueeze_dimq_typek_typeq_embedk_embedr$   r$   r%   apply_rotary_pos_embg   s   

rz   c                       s   e Zd ZdZded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j	eej	 f fddZ  ZS )OlmoAttentionz=Multi-headed attention from 'Attention Is All You Need' paperr;   	layer_idxc                    s   t    || _|| _t|d|j|j | _|j|j | _	| jd | _
|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 |j|jd| _d S )NrU   g      Tr9   )r   r   r;   r|   getattrr   num_attention_headsrU   rS   rc   r]   attention_dropout	is_causalr=   r>   attention_biasq_projk_projv_projo_projr!   r;   r|   r"   r$   r%   r      s(   
zOlmoAttention.__init__past_key_valuepast_key_values4.58new_nameversionNr&   position_embeddingsr\   cache_positionr   c                 K   s  |j d d }g |d| jR }| |}	| |}
| |}| jjd urJ|	j| jj | jjd |
j| jj | jjd |j| jj | jjd |	|	dd}	|
|	dd}
||	dd}|\}}t
|	|
||\}	}
|d ur|||d}||
|| j|\}
}t}| jjdkrt| jj }|| |	|
||f| jsdn| j| jd|\}}|jg |dR   }| |}||fS )	NrF   )minmaxr   rG   )rs   rr   r   eagerrW   )r^   r]   )rJ   rU   r   r   r   r;   clip_qkvclamp_viewre   rz   updater|   rn   _attn_implementationr   rb   r   r]   rQ   rh   r   )r!   r&   r   r\   r   r   r_   input_shapehidden_shapequery_statesri   rj   rr   rs   cache_kwargsattention_interfacerm   rk   r$   r$   r%   r/      sF   





zOlmoAttention.forward)NN)r0   r1   r2   r3   r   r4   r   r   r,   r5   tupler   r   
LongTensorr/   r6   r$   r$   r"   r%   r{      s&    r{   c                       s   e Zd Zdedef fddZedddd							
				ddejde	ej de	ej
 de	e de	e de	ej
 de	eejejf  dee dejfddZ  ZS )OlmoDecoderLayerr;   r|   c                    sF   t    |j| _t||d| _t|| _t|j| _t|j| _	d S )N)r;   r|   )
r   r   r   r{   	self_attnr7   mlpr   input_layernormpost_attention_layernormr   r"   r$   r%   r      s   

zOlmoDecoderLayer.__init__r   r   r   r   NFr&   r\   rt   	use_cacher   r   r_   r   c              
   K   s^   |}	|  |}| jd|||||||d|\}}
|	| }|}	| |}| |}|	| }|S )N)r&   r\   rt   r   r   r   r   r$   )r   r   r   r   )r!   r&   r\   rt   r   r   r   r   r_   residual_r$   r$   r%   r/      s&   




zOlmoDecoderLayer.forward)NNNFNN)r0   r1   r2   r   r4   r   r   r,   r5   r   r   r   boolr   r   r   r/   r6   r$   r$   r"   r%   r      s8    		
r   c                       sD   e Zd ZU ejed< ddef fddZe e	dd Z
  ZS )	OlmoRotaryEmbeddinginv_freqNr;   c                    s   t    t|drt|jtr|jd|jd| _nd| _|j| _	|j| _
|| _t| j | _| | j|\}| _| jd|dd | j| _d S )Nrope_scaling	rope_typetypedefaultr   F)
persistent)r   r   hasattr
isinstancer   dictgetr   max_position_embeddingsmax_seq_len_cachedoriginal_max_seq_lenr;   r   rope_init_fnattention_scalingregister_bufferr   original_inv_freq)r!   r;   devicer   r"   r$   r%   r     s   
zOlmoRotaryEmbedding.__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 }	||	fW  d    S 1 sqw   Y  d S )
Nr   rF   r   mpscpuF)device_typeenabledrG   rH   )r   floatrP   rJ   r+   r   r   r   strr,   autocastre   rK   rr   r   rs   )
r!   rE   rt   inv_freq_expandedposition_ids_expandedr   freqsembrr   rs   r$   r$   r%   r/     s   0&$zOlmoRotaryEmbedding.forwardr   )r0   r1   r2   r,   r5   __annotations__r   r   no_gradr   r/   r6   r$   r$   r"   r%   r      s   
 
r   c                   @   sH   e Zd ZU eed< dZdZdgZdgZdZ	dZ
dZdZdZeedZdS )OlmoPreTrainedModelr;   modelTr   r   )r&   
attentionsN)r0   r1   r2   r   r   base_model_prefixsupports_gradient_checkpointing_no_split_modules_skip_keys_device_placement_supports_flash_attn_supports_sdpa_supports_flex_attn_can_compile_fullgraph_supports_attention_backendr   r{   _can_record_outputsr$   r$   r$   r%   r   #  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 )	OlmoModelr;   c                    s|   t     j| _ j| _t j j| j| _t	 fddt
 jD | _t j| _t d| _d| _|   d S )Nc                    s   g | ]}t  |qS r$   )r   ).0r|   r;   r$   r%   
<listcomp>?  s    z&OlmoModel.__init__.<locals>.<listcomp>r   F)r   r   pad_token_idpadding_idx
vocab_sizer=   	Embeddingr   embed_tokens
ModuleListrangenum_hidden_layerslayersr   normr   
rotary_embgradient_checkpointing	post_initrD   r"   r   r%   r   8  s   zOlmoModel.__init__N	input_idsr\   rt   r   inputs_embedsr   r   r_   r   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   )r   )r;   input_embedsr\   r   r   rt   )r\   rt   r   r   r   )last_hidden_stater   )
ValueErrorr   r   r;   get_seq_lengthr,   arangerJ   r   ro   r
   r   r   r   r   r   )r!   r   r\   rt   r   r   r   r   r_   past_seen_tokensrl   r&   r   decoder_layerr$   r$   r%   r/   H  sP   

	

zOlmoModel.forward)NNNNNNN)r0   r1   r2   r   r   r   r   r   r,   r   r5   r   FloatTensorr   r   r   r   r/   r6   r$   r$   r"   r%   r   6  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 )OlmoForCausalLMzlm_head.weightlm_headcolwise_repr&   logitsc                    s@   t  | t|| _|j| _tj|j|jdd| _| 	  d S r8   )
r   r   r   r   r   r=   r>   r   r   r   rD   r"   r$   r%   r     s
   
zOlmoForCausalLM.__init__Nr   r   r\   rt   r   r   labelsr   r   logits_to_keepr_   r   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, OlmoForCausalLM

        >>> model = OlmoForCausalLM.from_pretrained("meta-olmo/Olmo-2-7b-hf")
        >>> tokenizer = AutoTokenizer.from_pretrained("meta-olmo/Olmo-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   r\   rt   r   r   r   r   N)r   r   r   )lossr   r   r&   r   r$   )r   r   r   r4   slicer   loss_functionr;   r   r   r   r&   r   )r!   r   r\   rt   r   r   r   r   r   r   r_   outputsr&   slice_indicesr   r   r$   r$   r%   r/     s0    zOlmoForCausalLM.forward)	NNNNNNNNr   )r0   r1   r2   _tied_weights_keys_tp_plan_pp_planr   r   r   r   r,   r   r5   r   r   r   r   r4   r   r   r   r/   r6   r$   r$   r"   r%   r     sN    		
r   )r   r   r   )rW   )Nr   );typingr   r   r   r,   torch.nnr=   torch.nn.functionalrf   r)   activationsr   cache_utilsr   r   
generationr	   masking_utilsr
   modeling_layersr   modeling_outputsr   r   modeling_rope_utilsr   r   modeling_utilsr   r   processing_utilsr   utilsr   r   r   utils.deprecationr   utils.genericr   configuration_olmor   Moduler   r7   rN   r5   r4   rV   r   rn   rz   r{   r   r   r   r   r   __all__r$   r$   r$   r%   <module>   sd   

P-#NK