o
    wi7                     @   s\  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 dd
lmZ ddlmZ ddlmZmZ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# ddl$m%Z%m&Z&m'Z'm(Z( ddl)m*Z* e(+e,Z-dd Z.dBddZ/dej0de1dej0fddZ2	dCdej3dej0dej0d ej0d!eej0 d"e4d#e4fd$d%Z5G d&d' d'ej3Z6ed(G d)d* d*ej3Z7e&G d+d, d,e!Z8G d-d. d.ej3Z9G d/d0 d0eZ:G d1d2 d2ej3Z;e&G d3d4 d4e8Z<G d5d6 d6ee%Z=e&G d7d8 d8e8eZ>e&d9d:G d;d< d<e8Z?e&G d=d> d>e8Z@e&G d?d@ d@e8ZAg dAZBdS )D    )CallableOptionalUnionN)nn   )ACT2FN)CacheDynamicCache)GenerationMixin)use_kernel_forward_from_hub)create_causal_mask!create_sliding_window_causal_mask)FlashAttentionKwargs)GradientCheckpointingLayer)BaseModelOutputWithPastCausalLMOutputWithPastQuestionAnsweringModelOutput SequenceClassifierOutputWithPastTokenClassifierOutput)ROPE_INIT_FUNCTIONSdynamic_rope_update)ALL_ATTENTION_FUNCTIONSPreTrainedModel)Unpack)
LossKwargsauto_docstringcan_return_tuplelogging   )SmolLM3Configc                 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)shapetorchcat)xx1x2 r*   i/home/ubuntu/sommelier/.venv/lib/python3.10/site-packages/transformers/models/smollm3/modeling_smollm3.pyrotate_half3   s   r,   c                 C   sD   | |}| |}| | 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.
    )	unsqueezer,   )qkcossinposition_idsunsqueeze_dimq_embedk_embedr*   r*   r+   apply_rotary_pos_emb:   s
   

r6   hidden_states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$   expandreshape)r7   r8   batchnum_key_value_headsslenhead_dimr*   r*   r+   	repeat_kvU   s
   0r@           modulequerykeyvalueattention_maskscalingdropoutc                 K   s   t || j}t || j}	t||dd| }
|d ur3|d d d d d d d |jd f }|
| }
tjj|
dtj	d
|j}
tjj|
|| jd}
t|
|	}|dd }||
fS )Nr!   r   r    )r#   dtype)ptrainingr   )r@   num_key_value_groupsr%   matmul	transposer$   r   
functionalsoftmaxfloat32torJ   rH   rL   
contiguous)rB   rC   rD   rE   rF   rG   rH   kwargs
key_statesvalue_statesattn_weightscausal_maskattn_outputr*   r*   r+   eager_attention_forwarda   s   
&r[   c                       s   e Zd ZdZdedef fddZ		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
ej e
e	ej  f fddZ  ZS )SmolLM3Attentionz=Multi-headed attention from 'Attention Is All You Need' paperconfig	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| _|j| | _|jr||j| dkr||j| _d S d | _d S )Nr?   g      Tbiassliding_attention)super__init__r]   r^   getattrhidden_sizenum_attention_headsr?   r=   rM   rG   attention_dropout	is_causalr   Linearattention_biasq_projk_projv_projo_projno_rope_layersuse_ropeuse_sliding_windowlayer_typessliding_windowselfr]   r^   	__class__r*   r+   rc   ~   s8   
zSmolLM3Attention.__init__Nr7   position_embeddingsrF   past_key_valuecache_positionrU   r9   c                 K   s*  |j d d }g |d| jR }| ||dd}	| ||dd}
| ||dd}| jrE|\}}t|	|
||\}	}
|d urXd|i}|	|
|| j
|\}
}t}| jjdkrft| jj }|| |	|
||f| jsrdn| j| j| jd|\}}|jg |dR   }| |}||fS )Nr    r   r!   rz   eagerrA   )rH   rG   rs   )r$   r?   rk   viewrO   rl   rm   rp   r6   updater^   r[   r]   _attn_implementationr   rL   rg   rG   rs   r;   rT   rn   )ru   r7   rx   rF   ry   rz   rU   input_shapehidden_shapequery_statesrV   rW   r0   r1   cache_kwargsattention_interfacerZ   rX   r*   r*   r+   forward   s<   		

zSmolLM3Attention.forward)NN)__name__
__module____qualname____doc__r   intrc   r%   Tensortupler   r   
LongTensorr   r   r   __classcell__r*   r*   rv   r+   r\   {   s(    #r\   RMSNormc                       s.   e Zd Zd fdd	Zdd Zdd Z  ZS )	SmolLM3RMSNormư>c                    s&   t    tt|| _|| _dS )z=
        SmolLM3RMSNorm is equivalent to T5LayerNorm
        N)rb   rc   r   	Parameterr%   onesweightvariance_epsilon)ru   re   epsrv   r*   r+   rc      s   

zSmolLM3RMSNorm.__init__c                 C   sJ   |j }|tj}|djddd}|t|| j  }| j|| S )Nr!   r    T)keepdim)	rJ   rS   r%   rR   powmeanrsqrtr   r   )ru   r7   input_dtypevariancer*   r*   r+   r      s
   zSmolLM3RMSNorm.forwardc                 C   s   t | jj d| j S )Nz, eps=)r   r   r$   r   ru   r*   r*   r+   
extra_repr   s   zSmolLM3RMSNorm.extra_repr)r   )r   r   r   rc   r   r   r   r*   r*   rv   r+   r      s    r   c                   @   sL   e Zd ZeZdZdZdgZdgZdZ	dZ
dZdZdZdZdZdZdd ZdS )SmolLM3PreTrainedModelmodelTSmolLM3DecoderLayerpast_key_valuesc                 C   s   | j j}t|tjr"|jjjd|d |jd ur |jj	  d S d S t|tj
rC|jjjd|d |jd urA|jj|j 	  d S d S t|trQ|jjd d S d S )NrA   )r   stdg      ?)r]   initializer_range
isinstancer   ri   r   datanormal_r`   zero_	Embeddingpadding_idxr   fill_)ru   rB   r   r*   r*   r+   _init_weights   s   


z$SmolLM3PreTrainedModel._init_weightsN)r   r   r   r   config_classbase_model_prefixsupports_gradient_checkpointing_no_split_modules_skip_keys_device_placement_supports_flash_attn_3_supports_flash_attn_2_supports_sdpa_supports_flex_attn_supports_cache_class_supports_quantized_cache_supports_static_cache_supports_attention_backendr   r*   r*   r*   r+   r      s    r   c                       s$   e Zd Z fddZdd Z  ZS )
SmolLM3MLPc                    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 NFr_   )rb   rc   r]   re   intermediate_sizer   ri   	gate_projup_proj	down_projr   
hidden_actact_fnru   r]   rv   r*   r+   rc      s   
zSmolLM3MLP.__init__c                 C   s$   |  | | || | }|S N)r   r   r   r   )ru   r'   r   r*   r*   r+   r     s    zSmolLM3MLP.forward)r   r   r   rc   r   r   r*   r*   rv   r+   r      s    
r   c                       s   e Zd Zdedef fddZ							ddejdeej d	eej	 d
ee
 dee dee deej	 deeejejf  dee deejeeejejf  f fddZ  ZS )r   r]   r^   c                    s^   t    |j| _t||d| _t|| _t|j|jd| _	t|j|jd| _
|j| | _d S )N)r]   r^   r   )rb   rc   re   r\   	self_attnr   mlpr   rms_norm_epsinput_layernormpost_attention_layernormrr   attention_typert   rv   r*   r+   rc     s   

zSmolLM3DecoderLayer.__init__NFr7   rF   r2   ry   output_attentions	use_cacherz   rx   rU   r9   c	                 K   st   |}
|  |}| jd||||||||d|	\}}|
| }|}
| |}| |}|
| }|f}|r8||f7 }|S )N)r7   rF   r2   ry   r   r   rz   rx   r*   )r   r   r   r   )ru   r7   rF   r2   ry   r   r   rz   rx   rU   residualself_attn_weightsoutputsr*   r*   r+   r     s.   
	



zSmolLM3DecoderLayer.forward)NNNFFNN)r   r   r   r   r   rc   r%   r   r   r   r   boolr   r   r   FloatTensorr   r   r*   r*   rv   r+   r     s<    	
r   c                       s8   e Zd Zddef fddZe edd Z  Z	S )SmolLM3RotaryEmbeddingNr]   c                    s   t    t|dr|jd u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defaultinv_freqF)
persistent)rb   rc   hasattrr   getr   max_position_embeddingsmax_seq_len_cachedoriginal_max_seq_lenr]   r   rope_init_fnattention_scalingregister_bufferr   original_inv_freq)ru   r]   devicer   rv   r*   r+   rc   C  s   
zSmolLM3RotaryEmbedding.__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!   r"   )rJ   )r   floatr:   r$   rS   r   r   r   strr%   autocastrO   r&   r0   r   r1   rJ   )
ru   r'   r2   inv_freq_expandedposition_ids_expandedr   freqsembr0   r1   r*   r*   r+   r   T  s   0&zSmolLM3RotaryEmbedding.forwardr   )
r   r   r   r   rc   r%   no_gradr   r   r   r*   r*   rv   r+   r   B  s
    r   c                       s   e Zd Zdef fddZdd Z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 de	e de	e de	e
j dee defddZ  ZS )SmolLM3Modelr]   c                    s   t     j| _ j| _t j j| j| _t	 fddt
 jD | _t j jd| _t d| _d| _d| jjv | _|   d S )Nc                    s   g | ]}t  |qS r*   )r   ).0r^   r]   r*   r+   
<listcomp>m  s    z)SmolLM3Model.__init__.<locals>.<listcomp>r   r   Fra   )rb   rc   pad_token_idr   
vocab_sizer   r   re   embed_tokens
ModuleListrangenum_hidden_layerslayersr   r   normr   
rotary_embgradient_checkpointingr]   rr   has_sliding_layers	post_initr   rv   r   r+   rc   f  s   zSmolLM3Model.__init__c                 C      | j S r   r   r   r*   r*   r+   get_input_embeddingsw     z!SmolLM3Model.get_input_embeddingsc                 C   
   || _ d S r   r   ru   rE   r*   r*   r+   set_input_embeddingsz     
z!SmolLM3Model.set_input_embeddingsN	input_idsrF   r2   r   inputs_embedsr   r   output_hidden_statesrz   flash_attn_kwargsr9   c
                 K   s  |d ur|n| j j}|d ur|n| j j}|d ur|n| j j}|d u |d uA r*td| jr9| jr9|r9td d}t	|t
d tfsFtd|d u rO| |}|rX|d u rXt }|	d u rt|d urd| nd}tj|||jd  |jd}	|d u r}|	d}t	| }ts| j |||	||d}d	tdi |i}| jrtdi ||d
< |}| ||}|rdnd }|rdnd }| jd | j j D ])}|r||f7 }||f||j |||||	|d|
}|d }|r||d f7 }q| |}|r||f7 }t||r|nd ||dS )Nz:You must specify exactly one of input_ids or inputs_embedszX`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`.FzBThe `past_key_values` should be either a `Cache` object or `None`.r   r   r   )r]   input_embedsrF   rz   r   r2   full_attentionra   r*   )rF   r2   ry   r   r   rz   rx   )last_hidden_stater   r7   
attentions)r]   r   r  r   
ValueErrorr   rL   loggerwarning_oncer   r   r   r   r	   get_seq_lengthr%   aranger$   r   r-   dictr   r   r   r   r   r   r   r   r   )ru   r  rF   r2   r   r  r   r   r  rz   r	  past_seen_tokenscausal_mask_mappingmask_kwargsr7   rx   all_hidden_statesall_self_attnsdecoder_layerlayer_outputsr*   r*   r+   r   }  s   



	


zSmolLM3Model.forward	NNNNNNNNN)r   r   r   r   rc   r   r  r   r   r   r%   r   r   r   r   r   r   r   r   r   r   r*   r*   rv   r+   r   d  sL    	
r   c                   @   s   e Zd ZdS )KwargsForCausalLMN)r   r   r   r*   r*   r*   r+   r    s    r  c                       s
  e Zd ZdgZddiZddgdgfiZ fddZdd	 Zd
d Zdd Z	dd Z
dd Zdd Zee											d%deej deej deej dee deej deej dee dee dee deej d eeejf d!ee d"efd#d$Z  ZS )&SmolLM3ForCausalLMzlm_head.weightlm_headcolwise_repr7   logitsc                    s@   t  | t|| _|j| _tj|j|jdd| _| 	  d S r   )
rb   rc   r   r   r   r   ri   re   r  r   r   rv   r*   r+   rc     s
   
zSmolLM3ForCausalLM.__init__c                 C      | j jS r   r   r   r   r*   r*   r+   r        z'SmolLM3ForCausalLM.get_input_embeddingsc                 C      || j _d S r   r#  r  r*   r*   r+   r       z'SmolLM3ForCausalLM.set_input_embeddingsc                 C   r   r   r  r   r*   r*   r+   get_output_embeddings  r  z(SmolLM3ForCausalLM.get_output_embeddingsc                 C   r  r   r'  )ru   new_embeddingsr*   r*   r+   set_output_embeddings  r  z(SmolLM3ForCausalLM.set_output_embeddingsc                 C   r  r   r   )ru   decoderr*   r*   r+   set_decoder  r  zSmolLM3ForCausalLM.set_decoderc                 C   r   r   r+  r   r*   r*   r+   get_decoder  r  zSmolLM3ForCausalLM.get_decoderNr   r  rF   r2   r   r  labelsr   r   r  rz   logits_to_keeprU   r9   c                 K   s   |dur|n| j j}|	dur|	n| j j}	| jd||||||||	|
d	|}|j}t|tr4t| dn|}| |dd|ddf }d}|durX| j	d||| j j
d|}t|||j|j|jdS )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, SmolLM3ForCausalLM

        >>> model = SmolLM3ForCausalLM.from_pretrained("meta-smollm3/SmolLM3-2-7b-hf")
        >>> tokenizer = AutoTokenizer.from_pretrained("meta-smollm3/SmolLM3-2-7b-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."
        ```N)	r  rF   r2   r   r  r   r   r  rz   )r!  r/  r   lossr!  r   r7   r  r*   )r]   r   r  r   r  r   r   slicer  loss_functionr   r   r   r7   r  )ru   r  rF   r2   r   r  r/  r   r   r  rz   r0  rU   r   r7   slice_indicesr!  r2  r*   r*   r+   r     s:   '
zSmolLM3ForCausalLM.forward)NNNNNNNNNNr   )r   r   r   _tied_weights_keys_tp_plan_pp_planrc   r   r  r(  r*  r-  r.  r   r   r   r%   r   r   r   r   r   r   r   r   r  r   r   r   r*   r*   rv   r+   r    sf    		
r  a  
    The SmolLM3 Model transformer with a sequence classification head on top (linear layer).

    [`SmolLM3ForSequenceClassification`] uses the last token in order to do the classification, as other causal models
    (e.g. GPT-2) do.

    Since it does classification on the last token, it requires to know the position of the last token. If a
    `pad_token_id` is defined in the configuration, it finds the last token that is not a padding token in each row. If
    no `pad_token_id` is defined, it simply takes the last value in each row of the batch. Since it cannot guess the
    padding tokens when `inputs_embeds` are passed instead of `input_ids`, it does the same (take the last value in
    each row of the batch).
    )custom_introc                          e Zd Z fddZdd Z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e defddZ  ZS ) SmolLM3ForSequenceClassificationc                    s@   t  | |j| _t|| _tj|j| jdd| _| 	  d S r   )
rb   rc   
num_labelsr   r   r   ri   re   scorer   r   rv   r*   r+   rc   l  s
   
z)SmolLM3ForSequenceClassification.__init__c                 C   r"  r   r#  r   r*   r*   r+   r   u  r$  z5SmolLM3ForSequenceClassification.get_input_embeddingsc                 C   r%  r   r#  r  r*   r*   r+   r  x  r&  z5SmolLM3ForSequenceClassification.set_input_embeddingsNr  rF   r2   r   r  r/  r   r   r  r9   c
              
   C   s(  | j ||||||||	d}
|
j}| |}|dur|jd }n|jd }| jjdu r2|dkr2td| jjdu r;d}n1|dur`|| jjk|jt	j
}t	j|jd |jt	j
d}|| d}nd}t| jj d |t	j||jd	|f }d}|dur| j|||| jd
}t|||
j|
j|
jdS )  
        labels (`torch.LongTensor` of shape `(batch_size,)`, *optional*):
            Labels for computing the sequence classification/regression loss. Indices should be in `[0, ...,
            config.num_labels - 1]`. If `config.num_labels == 1` a regression loss is computed (Mean-Square loss), If
            `config.num_labels > 1` a classification loss is computed (Cross-Entropy).
        rF   r2   r   r  r   r   r  Nr   r   z=Cannot handle batch sizes > 1 if no padding token is defined.r    )r   rJ   z will not detect padding tokens in `inputs_embeds`. Results may be unexpected if using padding tokens in conjunction with `inputs_embeds.`r
  )r!  r/  pooled_logitsr]   r1  )r   r  r=  r$   r]   r   r  rS   r   r%   int32r  argmaxr  r  rw   r   r4  r   r   r7   r  )ru   r  rF   r2   r   r  r/  r   r   r  transformer_outputsr7   r!  
batch_sizelast_non_pad_tokennon_pad_masktoken_indicesr@  r2  r*   r*   r+   r   {  sL   


z(SmolLM3ForSequenceClassification.forwardr  )r   r   r   rc   r   r  r   r   r   r%   r   r   r   r   r   r   r   r   r*   r*   rv   r+   r;  ]  sH    		
r;  c                       r:  )SmolLM3ForTokenClassificationc                    s|   t  | |j| _t|| _t|dd d ur|j}nt|dd d ur'|j}nd}t	|| _
t|j|j| _|   d S )Nclassifier_dropouthidden_dropoutg?)rb   rc   r<  r   r   rd   rI  rJ  r   DropoutrH   ri   re   r=  r   )ru   r]   rI  rv   r*   r+   rc     s   
z&SmolLM3ForTokenClassification.__init__c                 C   r"  r   r#  r   r*   r*   r+   r     r$  z2SmolLM3ForTokenClassification.get_input_embeddingsc                 C   r%  r   r#  r  r*   r*   r+   r    r&  z2SmolLM3ForTokenClassification.set_input_embeddingsNr  rF   r2   r   r  r/  r   r   r  r9   c
              
   C   sd   | j ||||||||	d}
|
j}| |}| |}d}|dur(| ||| j}t|||
j|
jdS )r>  r?  N)r2  r!  r7   r  )	r   r  rH   r=  r4  r]   r   r7   r  )ru   r  rF   r2   r   r  r/  r   r   r  r   sequence_outputr!  r2  r*   r*   r+   r     s,   


z%SmolLM3ForTokenClassification.forwardr  )r   r   r   rc   r   r  r   r   r   r%   r   r   r   r   r   r   r   r   r*   r*   rv   r+   rH    sH    	
rH  c                       s   e Zd ZdZ fddZdd Z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
j de	e de	e defddZ  ZS )SmolLM3ForQuestionAnsweringtransformerc                    s2   t  | t|| _t|jd| _|   d S )Nr!   )	rb   rc   r   rN  r   ri   re   
qa_outputsr   r   rv   r*   r+   rc     s   
z$SmolLM3ForQuestionAnswering.__init__c                 C   r"  r   rN  r   r   r*   r*   r+   r     r$  z0SmolLM3ForQuestionAnswering.get_input_embeddingsc                 C   r%  r   rP  r  r*   r*   r+   r    r&  z0SmolLM3ForQuestionAnswering.set_input_embeddingsNr  rF   r2   r   r  start_positionsend_positionsr   r  r9   c
              	   K   s   | j |||||||	d}|j}| |}|jddd\}}|d }|d }d }|d urA|d urA| j||||fi |
}t||||j|j	dS )N)rF   r2   r   r  r   r  r   r    r"   )r2  start_logits
end_logitsr7   r  )
rN  r  rO  splitsqueezerT   r4  r   r7   r  )ru   r  rF   r2   r   r  rQ  rR  r   r  rU   r   rL  r!  rS  rT  r2  r*   r*   r+   r     s0   

z#SmolLM3ForQuestionAnswering.forwardr  )r   r   r   r   rc   r   r  r   r   r   r%   r   r   r   r   r   r   r   r   r*   r*   rv   r+   rM    sJ    	
rM  )r   r   r  r;  rH  rM  )Nr   )rA   )Ctypingr   r   r   r%   r   activationsr   cache_utilsr   r	   
generationr
   integrationsr   masking_utilsr   r   modeling_flash_attention_utilsr   modeling_layersr   modeling_outputsr   r   r   r   r   modeling_rope_utilsr   r   modeling_utilsr   r   processing_utilsr   utilsr   r   r   r   configuration_smollm3r   
get_loggerr   r  r,   r6   r   r   r@   Moduler   r[   r\   r   r   r   r   r   r   r  r  r;  rH  rM  __all__r*   r*   r*   r+   <module>   sv   


N6" 	lVF>