U ,-e@szdZddlmZddlmZddlmZddlmZddl m Z e e Z dd d d ZGd d d eZGdddeZdS)z SqueezeBERT model configuration) OrderedDict)Mapping)PretrainedConfig) OnnxConfig)loggingzOhttps://huggingface.co/squeezebert/squeezebert-uncased/resolve/main/config.jsonzLhttps://huggingface.co/squeezebert/squeezebert-mnli/resolve/main/config.jsonzUhttps://huggingface.co/squeezebert/squeezebert-mnli-headless/resolve/main/config.json)zsqueezebert/squeezebert-uncasedzsqueezebert/squeezebert-mnliz%squeezebert/squeezebert-mnli-headlesscs*eZdZdZeZdZdfdd ZZS)SqueezeBertConfigah This is the configuration class to store the configuration of a [`SqueezeBertModel`]. It is used to instantiate a SqueezeBERT 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 SqueezeBERT [squeezebert/squeezebert-uncased](https://huggingface.co/squeezebert/squeezebert-uncased) 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 30522): Vocabulary size of the SqueezeBERT model. Defines the number of different tokens that can be represented by the `inputs_ids` passed when calling [`SqueezeBertModel`]. hidden_size (`int`, *optional*, defaults to 768): Dimensionality of the encoder layers and the pooler layer. num_hidden_layers (`int`, *optional*, defaults to 12): Number of hidden layers in the Transformer encoder. num_attention_heads (`int`, *optional*, defaults to 12): Number of attention heads for each attention layer in the Transformer encoder. intermediate_size (`int`, *optional*, defaults to 3072): Dimensionality of the "intermediate" (often named feed-forward) layer in the Transformer encoder. hidden_act (`str` or `Callable`, *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. hidden_dropout_prob (`float`, *optional*, defaults to 0.1): The dropout probability for all fully connected layers in the embeddings, encoder, and pooler. attention_probs_dropout_prob (`float`, *optional*, defaults to 0.1): The dropout ratio for the attention probabilities. 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 [`BertModel`] or [`TFBertModel`]. 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): pad_token_id (`int`, *optional*, defaults to 0): The ID of the token in the word embedding to use as padding. embedding_size (`int`, *optional*, defaults to 768): The dimension of the word embedding vectors. q_groups (`int`, *optional*, defaults to 4): The number of groups in Q layer. k_groups (`int`, *optional*, defaults to 4): The number of groups in K layer. v_groups (`int`, *optional*, defaults to 4): The number of groups in V layer. post_attention_groups (`int`, *optional*, defaults to 1): The number of groups in the first feed forward network layer. intermediate_groups (`int`, *optional*, defaults to 4): The number of groups in the second feed forward network layer. output_groups (`int`, *optional*, defaults to 4): The number of groups in the third feed forward network layer. Examples: ```python >>> from transformers import SqueezeBertConfig, SqueezeBertModel >>> # Initializing a SqueezeBERT configuration >>> configuration = SqueezeBertConfig() >>> # Initializing a model (with random weights) from the configuration above >>> model = SqueezeBertModel(configuration) >>> # Accessing the model configuration >>> configuration = model.config ``` Attributes: pretrained_config_archive_map (Dict[str, str]): A dictionary containing all the available pre-trained checkpoints. Z squeezebert:w gelu皙?{Gz?-q=rc stjfd| i|||_||_||_||_||_||_||_||_ | |_ | |_ | |_ | |_ ||_||_||_||_||_||_||_dS)N pad_token_id)super__init__ vocab_size hidden_sizenum_hidden_layersnum_attention_heads hidden_actintermediate_sizehidden_dropout_probattention_probs_dropout_probmax_position_embeddingstype_vocab_sizeinitializer_rangelayer_norm_epsembedding_sizeq_groupsk_groupsv_groupspost_attention_groupsintermediate_groups output_groups)selfrrrrrrrrr r!r"r#rr$r%r&r'r(r)r*kwargs __class__z/var/www/html/Darija-Ai-Train/env/lib/python3.8/site-packages/transformers/models/squeezebert/configuration_squeezebert.pyrss(zSqueezeBertConfig.__init__)r r r r r r rrrrrrrr rrrrrr) __name__ __module__ __qualname____doc__)SQUEEZEBERT_PRETRAINED_CONFIG_ARCHIVE_MAPZpretrained_config_archive_mapZ model_typer __classcell__r/r/r-r0r%s0Jrc@s.eZdZeeeeeeffdddZdS)SqueezeBertOnnxConfig)returncCs<|jdkrdddd}n ddd}td|fd|fd |fgS) Nzmultiple-choicebatchchoicesequence)rrr)rrZ input_idsZattention_maskZtoken_type_ids)taskr)r+Z dynamic_axisr/r/r0inputss  zSqueezeBertOnnxConfig.inputsN)r1r2r3propertyrstrintr=r/r/r/r0r7sr7N)r4 collectionsrtypingrZconfiguration_utilsrZonnxrutilsrZ get_loggerr1loggerr5rr7r/r/r/r0s       ~