o
    ZhI                     @   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G dd de
ZG dd deZddg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                       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
    ```Z
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_cacheZnum_hidden_layersscale_embeddingsuper__init__)selfr*   r+   r-   r,   r   r/   r.   r0   r6   r7   r8   r%   r4   r   r1   r2   r3   r5   r&   r9   r"   r#   r$   r'   r(   kwargs	__class__r)   f/var/www/auris/lib/python3.10/site-packages/transformers/models/blenderbot/configuration_blenderbot.pyr;   k   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__Z
model_typeZkeys_to_ignore_at_inferenceZattribute_mapr;   __classcell__r)   r)   r>   r@   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)   r@   rR      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 )NrH   rK   rU   rV   zpresent.rW   rX   )rY   r:   outputsr   rZ   r\   r]   )r<   Zcommon_outputsZnum_encoder_layersr_   ra   r>   r)   r@   rb      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 )Zdecoder_r)   ).0nameZtensorr)   r)   r@   
<dictcomp>   s    zZBlenderbotOnnxConfig._generate_dummy_inputs_for_default_and_seq2seq_lm.<locals>.<dictcomp>ACannot generate dummy past_keys inputs without PyTorch installed.r   rJ   rN   rP   dimr   r)   )I_generate_dummy_inputs_for_sequence_classification_and_question_answeringrZ   itemsdictr
   
ValueErrortorchshaper   _configr   catonesr\   r]   appendzeros)r<   rd   re   rf   rg   rh   Zencoder_inputsZdecoder_seq_lengthZdecoder_inputsr^   rs   rK   Zencoder_seq_lengthnum_encoder_attention_headsZnum_decoder_attention_headsZencoder_shapeZdecoder_past_lengthZdecoder_shaper_   r`   r)   r)   r@   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 )Nrl   r   rJ   rM   )dtyper!   rm   c                    s    g | ]}    fqS r)   )ry   )ri   r_   Z
past_shapers   r)   r@   
<listcomp>8  s    zMBlenderbotOnnxConfig._generate_dummy_inputs_for_causal_lm.<locals>.<listcomp>r   )ro   rZ   r
   rr   rs   rt   r\   r   ru   r   r|   rv   rw   r]   )r<   rd   re   rf   rg   rh   r^   rK   ZseqlenZpast_key_values_lengthr_   r`   rz   Z
mask_dtyper)   r}   r@   $_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   )Zfixed_dimensionZnum_token_to_add )Zreturn_tensors)r   r   Zdefault_fixed_batchZnum_special_tokens_to_addZdefault_fixed_sequencejoinZ	unk_tokenrq   )	r<   rd   re   rf   rg   rh   Ztoken_to_addZdummy_inputr^   r)   r)   r@   ro   >  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 )NrH   )re   rf   rg   rh   rT   )rY   r{   r   ro   )r<   rd   re   rf   rg   rh   r^   r)   r)   r@   generate_dummy_inputsY  s   




z*BlenderbotOnnxConfig.generate_dummy_inputsc                    s:   | j dv rt ||||}d S tt| ||||}d S )NrH   )rY   r:   _flatten_past_key_values_r   )r<   Zflattened_outputrj   idxtr>   r)   r@   r   r  s
   

z.BlenderbotOnnxConfig._flatten_past_key_values_inputs_or_outputsrS   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)rR   rb   z4direction must either be "inputs" or "outputs", but z
 was givenrR   r   ZpresentZpast_encoder_sequenceZpast_decoder_sequencerO   rK   rV   .z.decoder.keyz.decoder.valuez.encoder.keyz.encoder.value)rr   r\   r]   )	r<   r   rS   rj   r_   r`   rL   rQ   ra   r)   r)   r@   r[   z  s   
z/BlenderbotOnnxConfig.fill_with_past_key_values_)rc   rc   FN)rA   rB   rC   propertyr   strintrR   rb   r   boolr   r	   r   r{   r   ro   r   r   r[   rE   r)   r)   r>   r@   rF      s     ($

<

(



*rF   N)rD   collectionsr   typingr   r   r    r   Zconfiguration_utilsr   Z
file_utilsr	   r
   Zonnxr   r   r   Z
onnx.utilsr   utilsr   Z
get_loggerrA   loggerr   rF   __all__r)   r)   r)   r@   <module>   s   
 
 d