
    g]I                         d 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mZ ddlmZmZmZ dd	lmZ dd
lmZ  ej        e          Z G d de
          Z G d de          ZdS )zBlenderbot model configuration    )OrderedDict)AnyMappingOptional   )PreTrainedTokenizer)PretrainedConfig)
TensorTypeis_torch_available)
OnnxConfigOnnxConfigWithPastOnnxSeq2SeqConfigWithPast) compute_effective_axis_dimension)loggingc                   j     e Zd ZdZdZdgZdddZ	 	 	 	 	 	 	 	 	 	 	 	 	 	 	 	 	 	 	 	 	 	 	 	 	 d fd	Z xZS )BlenderbotConfiga  
    This is the configuration class to store the configuration of a [`BlenderbotModel`]. It is used to instantiate an
    Blenderbot model according to the specified arguments, defining the model architecture. Instantiating a
    configuration with the defaults will yield a similar configuration to that of the Blenderbot
    [facebook/blenderbot-3B](https://huggingface.co/facebook/blenderbot-3B) architecture.

    Configuration objects inherit from [`PretrainedConfig`] and can be used to control the model outputs. Read the
    documentation from [`PretrainedConfig`] for more information.


    Args:
        vocab_size (`int`, *optional*, defaults to 50265):
            Vocabulary size of the Blenderbot model. Defines the number of different tokens that can be represented by
            the `inputs_ids` passed when calling [`BlenderbotModel`] or [`TFBlenderbotModel`].
        d_model (`int`, *optional*, defaults to 1024):
            Dimensionality of the layers and the pooler layer.
        encoder_layers (`int`, *optional*, defaults to 12):
            Number of encoder layers.
        decoder_layers (`int`, *optional*, defaults to 12):
            Number of decoder layers.
        encoder_attention_heads (`int`, *optional*, defaults to 16):
            Number of attention heads for each attention layer in the Transformer encoder.
        decoder_attention_heads (`int`, *optional*, defaults to 16):
            Number of attention heads for each attention layer in the Transformer decoder.
        decoder_ffn_dim (`int`, *optional*, defaults to 4096):
            Dimensionality of the "intermediate" (often named feed-forward) layer in decoder.
        encoder_ffn_dim (`int`, *optional*, defaults to 4096):
            Dimensionality of the "intermediate" (often named feed-forward) layer in decoder.
        activation_function (`str` or `function`, *optional*, defaults to `"gelu"`):
            The non-linear activation function (function or string) in the encoder and pooler. If string, `"gelu"`,
            `"relu"`, `"silu"` and `"gelu_new"` are supported.
        dropout (`float`, *optional*, defaults to 0.1):
            The dropout probability for all fully connected layers in the embeddings, encoder, and pooler.
        attention_dropout (`float`, *optional*, defaults to 0.0):
            The dropout ratio for the attention probabilities.
        activation_dropout (`float`, *optional*, defaults to 0.0):
            The dropout ratio for activations inside the fully connected layer.
        max_position_embeddings (`int`, *optional*, defaults to 128):
            The maximum sequence length that this model might ever be used with. Typically set this to something large
            just in case (e.g., 512 or 1024 or 2048).
        init_std (`float`, *optional*, defaults to 0.02):
            The standard deviation of the truncated_normal_initializer for initializing all weight matrices.
        encoder_layerdrop (`float`, *optional*, defaults to 0.0):
            The LayerDrop probability for the encoder. See the [LayerDrop paper](see https://arxiv.org/abs/1909.11556)
            for more details.
        decoder_layerdrop (`float`, *optional*, defaults to 0.0):
            The LayerDrop probability for the decoder. See the [LayerDrop paper](see https://arxiv.org/abs/1909.11556)
            for more details.
        scale_embedding (`bool`, *optional*, defaults to `False`):
            Scale embeddings by diving by sqrt(d_model).
        use_cache (`bool`, *optional*, defaults to `True`):
            Whether or not the model should return the last key/values attentions (not used by all models)
        forced_eos_token_id (`int`, *optional*, defaults to 2):
            The id of the token to force as the last generated token when `max_length` is reached. Usually set to
            `eos_token_id`.

    Example:

    ```python
    >>> from transformers import BlenderbotConfig, BlenderbotModel

    >>> # Initializing a Blenderbot facebook/blenderbot-3B style configuration
    >>> configuration = BlenderbotConfig()

    >>> # Initializing a model (with random weights) from the facebook/blenderbot-3B style configuration
    >>> model = BlenderbotModel(configuration)

    >>> # Accessing the model configuration
    >>> configuration = model.config
    ```
blenderbotpast_key_valuesencoder_attention_headsd_model)num_attention_headshidden_sizeH         (                 Tgelu 
  皙?{Gz?   Fr   r   c                 T   || _         || _        || _        || _        || _        || _        || _        || _        || _        || _	        || _
        || _        || _        || _        |	| _        |
| _        || _        || _        || _         t'                      j        d|||||||d| d S )N)pad_token_idbos_token_ideos_token_idis_encoder_decoderdecoder_start_token_idencoder_no_repeat_ngram_sizeforced_eos_token_id )
vocab_sizemax_position_embeddingsr   encoder_ffn_dimencoder_layersr   decoder_ffn_dimdecoder_layersdecoder_attention_headsdropoutattention_dropoutactivation_dropoutactivation_functioninit_stdencoder_layerdropdecoder_layerdrop	use_cachenum_hidden_layersscale_embeddingsuper__init__)selfr.   r/   r1   r0   r   r3   r2   r4   r:   r;   r<   r)   r8   r   r5   r6   r7   r9   r*   r>   r&   r'   r(   r+   r,   kwargs	__class__s                              s/var/www/html/ai-engine/env/lib/python3.11/site-packages/transformers/models/blenderbot/configuration_blenderbot.pyr@   zBlenderbotConfig.__init__k   s    : %'>$.,'>$.,'>$!2"4#6  !2!2"!/. 		
%%%1#9)E 3		
 		
 		
 		
 		
 		
 		
    )r   r   r   r   r   r   r   r   r   r   TTr    r!   r"   r   r   r#   r$   Fr   r$   r   r   r   )	__name__
__module____qualname____doc__
model_typekeys_to_ignore_at_inferenceattribute_mapr@   __classcell__rC   s   @rD   r   r      s        E EN J#4"5,EV_``M  # " "" %&5:
 :
 :
 :
 :
 :
 :
 :
 :
 :
rE   r   c                       e Zd Zedeeeeef         f         fd            Zedeeeeef         f         f fd            Z	 	 	 	 dde	ded	ed
e
dee         deeef         fdZ	 	 	 	 dde	ded	ed
e
dee         deeef         fdZ	 	 	 	 dde	ded	ed
e
dee         deeef         fdZ	 	 	 	 dde	ded	ed
e
dee         deeef         fdZ fdZdeeeeef         f         defdZ xZS )BlenderbotOnnxConfigreturnc           	         | j         dv rat          ddddfddddfg          }| j        rddi|d<   dd	d|d
<   nddd|d<   ddd|d
<   | j        r|                     |d           n| j         dk    rWt          ddddfddddfg          }| j        r4| j        \  }}t          |          D ]}ddd|d| d<   ddd|d| d<   n't          ddddfddddfddddfd
dddfg          }|S )Ndefaultz
seq2seq-lm	input_idsbatchencoder_sequence)r   r$   attention_maskr   decoder_input_ids past_decoder_sequence + sequencedecoder_attention_maskdecoder_sequenceinputs)	direction	causal-lmpast_sequence + sequencer   r   zpast_key_values..key.value)taskr   use_pastfill_with_past_key_values_
num_layersrange)rA   common_inputs_num_decoder_layersis        rD   r]   zBlenderbotOnnxConfig.inputs   s   9111' g2D"E"EF%77I'J'JK M } ^67\12>EJl:m:m6779@EW5X5X12>EJ\:]:]67} S///RRRY+%%' g2D"E"EF%77I'J'JK M } n(,%%122 n nADKPj@k@kM"<Q"<"<"<=FMRlBmBmM">Q">">">??' g2D"E"EF%77I'J'JK(g:L*M*MN-7?Q/R/RS	 M rE   c                     | j         dv rt                      j        }nUt          t          |           j        }| j        r4| j        \  }}t          |          D ]}ddd|d| d<   ddd|d| d<   |S )NrS   rV   r`   ra   zpresent.rb   rc   )rd   r?   outputsr   re   rg   rh   )rA   common_outputsnum_encoder_layersrj   rl   rC   s        rD   rn   zBlenderbotOnnxConfig.outputs   s     9111"WW_NN"#5t<<DN} g(,%"A122 g gA=DIc9d9dN#5a#5#5#56?FKe;f;fN#7a#7#7#788rE   FN	tokenizer
batch_size
seq_lengthis_pair	frameworkc           	      X   |                      |||||          }| j        s|nd}|                      |||||          }d |                                D             }t          d
i ||}	| j        r8t	                      st          d          dd l}
|	d         j        \  }}|	d         j        d         }| j        \  }}|||| j	        j
        |z  f}|}|||| j	        j
        |z  f}|
                    |	d         |
                    ||          gd          |	d<   g |	d	<   | j        \  }}t          |          D ]m}|	d	                             |
                    |          |
                    |          |
                    |          |
                    |          f           n|	S )Nr$   c                      i | ]\  }}d | |S )decoder_r-   ).0nametensors      rD   
<dictcomp>zZBlenderbotOnnxConfig._generate_dummy_inputs_for_default_and_seq2seq_lm.<locals>.<dictcomp>   s'    ___f+T++V___rE   ACannot generate dummy past_keys inputs without PyTorch installed.r   rU   rY   r[   dimr   r-   )I_generate_dummy_inputs_for_sequence_classification_and_question_answeringre   itemsdictr   
ValueErrortorchshaper   _configr   catonesrg   rh   appendzeros)rA   rr   rs   rt   ru   rv   encoder_inputsdecoder_seq_lengthdecoder_inputsri   r   rV   encoder_seq_lengthnum_encoder_attention_headsnum_decoder_attention_headsencoder_shapedecoder_past_lengthdecoder_shaperj   rk   s                       rD   1_generate_dummy_inputs_for_default_and_seq2seq_lmzFBlenderbotOnnxConfig._generate_dummy_inputs_for_default_and_seq2seq_lm   s    ggz:w	
 
 04}CZZ!ggz#5w	
 
 `_H\H\H^H^___@@~@@@= #	%''  !deee(5k(B(H%E%!./B!C!I!!LGKG_D')D+"(,GG	M #5+#(,GG	M 7<ii78%**UL_:`:`agh 7@ 7 7M23 02M+,$(O!A!-..  /077M22M22M22M22	    rE   c                    |                      |||||          }| j        rt                      st          d          dd l|d         j        \  }}|}	| j        \  }
}| j        \  }}
|||	| j        j	        |z  f|d         j
        }                    |d                             ||	|          gd          |d<   fdt          |          D             |d	<   |S )
Nr~   r   rU   rX   )dtyper$   r   c                 d    g | ],}                                                              f-S r-   )r   )rz   rj   
past_shaper   s     rD   
<listcomp>zMBlenderbotOnnxConfig._generate_dummy_inputs_for_causal_lm.<locals>.<listcomp>8  sC     0 0 0GHZ((%++j*A*AB0 0 0rE   r   )r   re   r   r   r   r   rg   r   r   r   r   r   r   rh   )rA   rr   rs   rt   ru   rv   ri   rV   seqlenpast_key_values_lengthrj   rk   r   
mask_dtyper   r   s                 @@rD   $_generate_dummy_inputs_for_causal_lmz9BlenderbotOnnxConfig._generate_dummy_inputs_for_causal_lm  s?    ffz:w	
 
 = 	%''  !deee)+6<ME6%+"$(O!A!-1-E*'+&(,GG	J ''78>J.3ii/0%**UDZbl*2m2mntu /8 / /M*+0 0 0 0 0LQRdLeLe0 0 0M+, rE   c                    t          |t          j        d          }|                    |          }t          |t          j        |          }d                    |j        g          |z  g|z  }t           |||                    }|S )Nr   )fixed_dimensionnum_token_to_add )return_tensors)r   r   default_fixed_batchnum_special_tokens_to_adddefault_fixed_sequencejoin	unk_tokenr   )	rA   rr   rs   rt   ru   rv   token_to_adddummy_inputri   s	            rD   r   z^BlenderbotOnnxConfig._generate_dummy_inputs_for_sequence_classification_and_question_answering>  s     6
(FYZ
 
 


 !::7CC5
(I\h
 
 


 xx!4 566CDzQYY{9MMMNNrE   c                     | j         dv r|                     |||||          }n@| j         dk    r|                     |||||          }n|                     |||||          }|S )NrS   )rs   rt   ru   rv   r_   )rd   r   r   r   )rA   rr   rs   rt   ru   rv   ri   s          rD   generate_dummy_inputsz*BlenderbotOnnxConfig.generate_dummy_inputsY  s     9111 RRjZQXdm S  MM Y+%% EEjZQXdm F  MM !jjjZQXdm k  M rE   c                     | j         dv r&t                                          ||||          }d S t          t          |                               ||||          }d S )NrS   )rd   r?   _flatten_past_key_values_r   )rA   flattened_outputr{   idxtrC   s        rD   r   z.BlenderbotOnnxConfig._flatten_past_key_values_r  si    9111$ww@@AQSWY\^_``$%>EE__ $Q   rE   inputs_or_outputsr^   c                    |dvrt          d| d          |dk    rdnd}| j        \  }}d}|dk    rdnd	}t          |          D ]:}d
|d|| d| d<   d
|d|| d| d<   d
|d|| d| d<   d
|d|| d| d<   ;d S )N)r]   rn   z4direction must either be "inputs" or "outputs", but z
 was givenr]   r   presentpast_encoder_sequencepast_decoder_sequencerZ   rV   ra   .z.decoder.keyz.decoder.valuez.encoder.keyz.encoder.value)r   rg   rh   )	rA   r   r^   r{   rj   rk   rW   r\   rl   s	            rD   rf   z/BlenderbotOnnxConfig.fill_with_past_key_values_z  s%   111iT]iiijjj$-$9$9  y $26?86K6K22Qs)** 	_ 	_A?FK[;\;\777778AHM]=^=^99999:?FK[;\;\777778AHM]=^=^99999::		_ 	_rE   )rq   rq   FN)rF   rG   rH   propertyr   strintr]   rn   r   boolr   r
   r   r   r   r   r   r   rf   rM   rN   s   @rD   rP   rP      s       &WS#X%6 67 & & & X&P 
gc3h&7!78 
 
 
 
 
 X
 *.7 7&7 7 	7
 7 J'7 
c	7 7 7 7x *." "&" " 	"
 " J'" 
c	" " " "P *. &  	
  J' 
c	   < *. &  	
  J' 
c	   2    _GCQTVYQYIZDZ<[ _hk _ _ _ _ _ _ _ _rE   rP   N)rI   collectionsr   typingr   r   r    r   configuration_utilsr	   
file_utilsr
   r   onnxr   r   r   
onnx.utilsr   utilsr   
get_loggerrF   loggerr   rP   r-   rE   rD   <module>r      sL   % $ # # # # # # ) ) ) ) ) ) ) ) ) ) # # # # # # 3 3 3 3 3 3 8 8 8 8 8 8 8 8 M M M M M M M M M M : : : : : :       
	H	%	%F
 F
 F
 F
 F
' F
 F
 F
R`_ `_ `_ `_ `_4 `_ `_ `_ `_ `_rE   