o
    hIJ                     @   s   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eZddiZG dd de
ZG d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)loggingzfacebook/blenderbot-3BzFhttps://huggingface.co/facebook/blenderbot-3B/resolve/main/config.jsonc                       sh   e Zd ZdZdZdgZdddZ					
			
																				d fdd	Z  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              
      s   || _ || _|| _|| _|| _|| _|| _|| _|| _|| _	|| _
|| _|| _|| _|	| _|
| _|| _|| _|| _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,   r.   r-   r   r0   r/   r1   r7   r8   r9   r&   r5   r   r2   r3   r4   r6   r'   r;   r#   r$   r%   r(   r)   kwargs	__class__r*   m/var/www/html/ai/venv/lib/python3.10/site-packages/transformers/models/blenderbot/configuration_blenderbot.pyr=   p   s<   
zBlenderbotConfig.__init__)r   r   r   r   r   r   r   r   r   r   TTr   r   r    r   r   r!   r"   Fr   r"   r   r   r   )	__name__
__module____qualname____doc__
model_typekeys_to_ignore_at_inferenceattribute_mapr=   __classcell__r*   r*   r@   rB   r   $   s>    G
r   c                       sf  e Zd Zedeeeeef f fddZedeeeeef f f fddZ				dd	e	d
edede
dee deeef fddZ				dd	e	d
edede
dee deeef fddZ				dd	e	d
edede
dee deeef fddZ				dd	e	d
edede
dee deeef fddZ fddZdeeeeef f defddZ  ZS )BlenderbotOnnxConfigreturnc                 C   s4  | j dv r@t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>| j|dd |S | j dkr|tddddfddddfg}| jrz| j\}}t|D ]}ddd|d| d< ddd|d| d< qa|S 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)r>   common_inputs_num_decoder_layersir*   r*   rB   rW      sD   


	zBlenderbotOnnxConfig.inputsc                    sp   | j dv rt j}|S tt| j}| jr6| j\}}t|D ]}ddd|d| d< ddd|d| d< q|S )NrM   rP   rZ   r[   zpresent.r\   r]   )r^   r<   outputsr   r_   ra   rb   )r>   common_outputsnum_encoder_layersrd   rf   r@   r*   rB   rg      s   

zBlenderbotOnnxConfig.outputsFN	tokenizer
batch_size
seq_lengthis_pair	frameworkc              	   C   s8  |  |||||}| js|nd}|  |||||}dd | D }tdi ||}	| jrt s5tddd l}
|	d j\}}|	d jd }| j\}}|||| j	j
| f}|}|||| j	j
| f}|
j|	d |
||gdd	|	d< g |	d
< | j\}}t|D ]}|	d
 |
||
||
||
|f q|	S )Nr"   c                 S   s   i | ]
\}}d | |qS )decoder_r*   ).0nametensorr*   r*   rB   
<dictcomp>   s    zZBlenderbotOnnxConfig._generate_dummy_inputs_for_default_and_seq2seq_lm.<locals>.<dictcomp>ACannot generate dummy past_keys inputs without PyTorch installed.r   rO   rS   rU   dimr   r*   )I_generate_dummy_inputs_for_sequence_classification_and_question_answeringr_   itemsdictr
   
ValueErrortorchshaper   _configr   catonesra   rb   appendzeros)r>   rk   rl   rm   rn   ro   encoder_inputsdecoder_seq_lengthdecoder_inputsrc   r|   rP   encoder_seq_lengthnum_encoder_attention_headsnum_decoder_attention_headsencoder_shapedecoder_past_lengthdecoder_shaperd   re   r*   r*   rB   1_generate_dummy_inputs_for_default_and_seq2seq_lm   sR   






zFBlenderbotOnnxConfig._generate_dummy_inputs_for_default_and_seq2seq_lmc                    s   |  |||||}| jrZt stddd l|d j\}}|}	| j\}
}| j\}}
|||	| jj	| f |d j
}j|d j||	|dgdd|d<  fdd	t|D |d
< |S )Nru   r   rO   rR   )dtyper"   rv   c                    s    g | ]}    fqS r*   )r   )rq   rd   
past_shaper|   r*   rB   
<listcomp>=  s    zMBlenderbotOnnxConfig._generate_dummy_inputs_for_causal_lm.<locals>.<listcomp>r   )rx   r_   r
   r{   r|   r}   ra   r   r~   r   r   r   r   rb   )r>   rk   rl   rm   rn   ro   rc   rP   seqlenpast_key_values_lengthrd   re   r   
mask_dtyper*   r   rB   $_generate_dummy_inputs_for_causal_lm  s0   






z9BlenderbotOnnxConfig._generate_dummy_inputs_for_causal_lmc           	      C   sV   t |tjdd}||}t |tj|d}d|jg| g| }t|||d}|S )Nr   )fixed_dimensionnum_token_to_add )return_tensors)r   r   default_fixed_batchnum_special_tokens_to_adddefault_fixed_sequencejoin	unk_tokenrz   )	r>   rk   rl   rm   rn   ro   token_to_adddummy_inputrc   r*   r*   rB   rx   C  s   
z^BlenderbotOnnxConfig._generate_dummy_inputs_for_sequence_classification_and_question_answeringc                 C   s\   | j dv r| j|||||d}|S | j dkr"| j|||||d}|S | j|||||d}|S )NrM   )rl   rm   rn   ro   rY   )r^   r   r   rx   )r>   rk   rl   rm   rn   ro   rc   r*   r*   rB   generate_dummy_inputs^  s   




z*BlenderbotOnnxConfig.generate_dummy_inputsc                    s:   | j dv rt ||||}d S tt| ||||}d S )NrM   )r^   r<   _flatten_past_key_values_r   )r>   flattened_outputrr   idxtr@   r*   rB   r   w  s
   

z.BlenderbotOnnxConfig._flatten_past_key_values_inputs_or_outputsrX   c           	      C   s   |dvrt d| d|dkrdnd}| j\}}d}|dkr!dnd	}t|D ]6}d
|d|| d| d< d
|d|| d| d< d
|d|| d| d< d
|d|| d| d< q'd S )N)rW   rg   z4direction must either be "inputs" or "outputs", but z
 was givenrW   r   presentpast_encoder_sequencepast_decoder_sequencerT   rP   r[   .z.decoder.keyz.decoder.valuez.encoder.keyz.encoder.value)r{   ra   rb   )	r>   r   rX   rr   rd   re   rQ   rV   rf   r*   r*   rB   r`     s   
z/BlenderbotOnnxConfig.fill_with_past_key_values_)rj   rj   FN)rC   rD   rE   propertyr   strintrW   rg   r   boolr   r	   r   r   r   rx   r   r   r`   rJ   r*   r*   r@   rB   rK      s     ($

<

(



*rK   N)rF   collectionsr   typingr   r   r    r   configuration_utilsr   
file_utilsr	   r
   onnxr   r   r   
onnx.utilsr   utilsr   
get_loggerrC   logger(BLENDERBOT_PRETRAINED_CONFIG_ARCHIVE_MAPr   rK   r*   r*   r*   rB   <module>   s   
 
