U 5AfH@sndZddlmZddlmZddlmZddlmZddl m Z e e Z Gdd d eZGd d d eZd S) zRemBERT model configuration) OrderedDict)Mapping)PretrainedConfig) OnnxConfig)loggingcs&eZdZdZdZdfdd ZZS) RemBertConfiga This is the configuration class to store the configuration of a [`RemBertModel`]. It is used to instantiate an RemBERT 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 RemBERT [google/rembert](https://huggingface.co/google/rembert) 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 250300): Vocabulary size of the RemBERT model. Defines the number of different tokens that can be represented by the `inputs_ids` passed when calling [`RemBertModel`] or [`TFRemBertModel`]. Vocabulary size of the model. Defines the different tokens that can be represented by the *inputs_ids* passed to the forward method of [`RemBertModel`]. hidden_size (`int`, *optional*, defaults to 1152): Dimensionality of the encoder layers and the pooler layer. num_hidden_layers (`int`, *optional*, defaults to 32): Number of hidden layers in the Transformer encoder. num_attention_heads (`int`, *optional*, defaults to 18): Number of attention heads for each attention layer in the Transformer encoder. input_embedding_size (`int`, *optional*, defaults to 256): Dimensionality of the input embeddings. output_embedding_size (`int`, *optional*, defaults to 1664): Dimensionality of the output embeddings. intermediate_size (`int`, *optional*, defaults to 4608): Dimensionality of the "intermediate" (i.e., feed-forward) layer in the Transformer encoder. hidden_act (`str` or `function`, *optional*, defaults to `"gelu"`): The non-linear activation function (function or string) in the encoder and pooler. If string, `"gelu"`, `"relu"`, `"selu"` and `"gelu_new"` are supported. hidden_dropout_prob (`float`, *optional*, defaults to 0): The dropout probability for all fully connected layers in the embeddings, encoder, and pooler. attention_probs_dropout_prob (`float`, *optional*, defaults to 0): The dropout ratio for the attention probabilities. classifier_dropout_prob (`float`, *optional*, defaults to 0.1): The dropout ratio for the classifier layer when fine-tuning. max_position_embeddings (`int`, *optional*, defaults to 512): 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). type_vocab_size (`int`, *optional*, defaults to 2): The vocabulary size of the `token_type_ids` passed when calling [`RemBertModel`] or [`TFRemBertModel`]. initializer_range (`float`, *optional*, defaults to 0.02): The standard deviation of the truncated_normal_initializer for initializing all weight matrices. layer_norm_eps (`float`, *optional*, defaults to 1e-12): The epsilon used by the layer normalization layers. is_decoder (`bool`, *optional*, defaults to `False`): Whether the model is used as a decoder or not. If `False`, the model is used as an encoder. use_cache (`bool`, *optional*, defaults to `True`): Whether or not the model should return the last key/values attentions (not used by all models). Only relevant if `config.is_decoder=True`. Example: ```python >>> from transformers import RemBertModel, RemBertConfig >>> # Initializing a RemBERT rembert style configuration >>> configuration = RemBertConfig() >>> # Initializing a model from the rembert style configuration >>> model = RemBertModel(configuration) >>> # Accessing the model configuration >>> configuration = model.config ```Zrembert gelu皙?{Gz?-q=Tr89c stjf|||d|||_||_||_| |_||_||_||_||_ ||_ | |_ | |_ | |_ ||_| |_||_||_d|_dS)N) pad_token_id bos_token_id eos_token_idF)super__init__ vocab_sizeinput_embedding_sizeoutput_embedding_sizemax_position_embeddings hidden_sizenum_hidden_layersnum_attention_headsintermediate_size hidden_acthidden_dropout_probattention_probs_dropout_probclassifier_dropout_probinitializer_rangetype_vocab_sizelayer_norm_eps use_cacheZtie_word_embeddings)selfrr"r#r$rr r%r&r'r(r)r!r+r*r,r-rrrkwargs __class__U/tmp/pip-unpacked-wheel-zw5xktn0/transformers/models/rembert/configuration_rembert.pyrbs$zRemBertConfig.__init__)r r r r r rrrrrrrrrrTrrr)__name__ __module__ __qualname____doc__Z model_typer __classcell__r2r2r0r3rs,Crc@s@eZdZeeeeeeffdddZeedddZ dS)RemBertOnnxConfig)returncCs<|jdkrdddd}n ddd}td|fd|fd |fgS) Nzmultiple-choicebatchchoicesequence)rr)rr>Z input_idsZattention_maskZtoken_type_ids)Ztaskr)r.Z dynamic_axisr2r2r3inputss  zRemBertOnnxConfig.inputscCsdS)Ng-C6?r2)r.r2r2r3atol_for_validationsz%RemBertOnnxConfig.atol_for_validationN) r4r5r6propertyrstrintr?floatr@r2r2r2r3r9s r9N)r7 collectionsrtypingrZconfiguration_utilsrZonnxrutilsrZ get_loggerr4loggerrr9r2r2r2r3s      r