qi#ldZddlmZddlmZddlmZejeZ GddeZ dgZ y)zGraniteMoe model configuration)PreTrainedConfig)RopeParameters)loggingc6xeZdZdZdZdgZ d dedzdedzdedzdedzd edzd edzd edzd edzd edzdedzde dzdedzdedzdedzde dzde e ee fzdzde dzdedzdedzdedzdedzdedzdedzdedzde dzdedzf4fd Z xZ S)!GraniteMoeConfiga This is the configuration class to store the configuration of a [`GraniteMoeModel`]. It is used to instantiate an GraniteMoe 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 GraniteMoe-3B. 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 32000): Vocabulary size of the GraniteMoe model. Defines the number of different tokens that can be represented by the `inputs_ids` passed when calling [`GraniteMoeModel`] hidden_size (`int`, *optional*, defaults to 4096): Dimension of the hidden representations. intermediate_size (`int`, *optional*, defaults to 11008): Dimension of the MLP representations. num_hidden_layers (`int`, *optional*, defaults to 32): Number of hidden layers in the Transformer decoder. num_attention_heads (`int`, *optional*, defaults to 32): Number of attention heads for each attention layer in the Transformer decoder. num_key_value_heads (`int`, *optional*): This is the number of key_value heads that should be used to implement Grouped Query Attention. If `num_key_value_heads=num_attention_heads`, the model will use Multi Head Attention (MHA), if `num_key_value_heads=1` the model will use Multi Query Attention (MQA) otherwise GQA is used. When converting a multi-head checkpoint to a GQA checkpoint, each group key and value head should be constructed by meanpooling all the original heads within that group. For more details, check out [this paper](https://huggingface.co/papers/2305.13245). If it is not specified, will default to `num_attention_heads`. hidden_act (`str` or `function`, *optional*, defaults to `"silu"`): The non-linear activation function (function or string) in the decoder. max_position_embeddings (`int`, *optional*, defaults to 2048): The maximum sequence length that this model might ever be used with. initializer_range (`float`, *optional*, defaults to 0.02): The standard deviation of the truncated_normal_initializer for initializing all weight matrices. rms_norm_eps (`float`, *optional*, defaults to 1e-06): The epsilon used by the rms normalization layers. 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`. pad_token_id (`int`, *optional*): Padding token id. bos_token_id (`int`, *optional*, defaults to 1): Beginning of stream token id. eos_token_id (`int`, *optional*, defaults to 2): End of stream token id. tie_word_embeddings (`bool`, *optional*, defaults to `False`): Whether to tie weight embeddings rope_parameters (`RopeParameters`, *optional*): Dictionary containing the configuration parameters for the RoPE embeddings. The dictionary should contain a value for `rope_theta` and optionally parameters used for scaling in case you want to use RoPE with longer `max_position_embeddings`. attention_bias (`bool`, *optional*, defaults to `False`): Whether to use a bias in the query, key, value and output projection layers during self-attention. attention_dropout (`float`, *optional*, defaults to 0.0): The dropout ratio for the attention probabilities. embedding_multiplier (`float`, *optional*, defaults to 1.0): embedding multiplier logits_scaling (`float`, *optional*, defaults to 1.0): divisor for output logits residual_multiplier (`float`, *optional*, defaults to 1.0): residual multiplier attention_multiplier (`float`, *optional*, defaults to 1.0): attention multiplier num_local_experts (`int`, *optional*, defaults to 8): total number of experts num_experts_per_tok (`int`, *optional*, defaults to 2): number of experts per token output_router_logits (`bool`, *optional*, defaults to `False`): Whether or not the router logits should be returned by the model. Enabling this will also allow the model to output the auxiliary loss. router_aux_loss_coef (`float`, *optional*, defaults to 0.001): router auxiliary loss coefficient ```python >>> from transformers import GraniteMoeModel, GraniteMoeConfig >>> # Initializing a GraniteMoe granitemoe-3b style configuration >>> configuration = GraniteMoeConfig() >>> # Initializing a model from the granitemoe-7b style configuration >>> model = GraniteMoeModel(configuration) >>> # Accessing the model configuration >>> configuration = model.config ``` granitemoepast_key_valuesN vocab_size hidden_sizeintermediate_sizenum_hidden_layersnum_attention_headsnum_key_value_heads hidden_actmax_position_embeddingsinitializer_range rms_norm_eps use_cache pad_token_id bos_token_id eos_token_idtie_word_embeddingsrope_parametersattention_biasattention_dropoutembedding_multiplierlogits_scalingresidual_multiplierattention_multipliernum_local_expertsnum_experts_per_tokoutput_router_logitsrouter_aux_loss_coefc ||_||_||_||_||_||_||}||_||_| |_| |_ | |_ ||_ ||_ ||_ ||_||_||_||_||_||_||_||_| |_| |_||_||_t5|ldi|y)N)r rr r r rrrrrrrrrrrrr r!r"r#rrrrrsuper__init__)selfr r r r rrrrrrrrrrrrrrrrrrr r!r"r#kwargs __class__s i/opt/pipecat/venv/lib/python3.12/site-packages/transformers/models/granitemoe/configuration_granitemoe.pyr'zGraniteMoeConfig.__init__qs<%'>$&!2!2#6   &"5 #6 $!2(",!2$8!,#6 $8!!2#6 $8!$8!#6 (((. "6")i}ii+ r-Nsiluig{Gz?gư>TNFNFg?r1r1r1r0FgMbP?)__name__ __module__ __qualname____doc__ model_typekeys_to_ignore_at_inferenceintstrfloatboolrdictr' __classcell__)r*s@r+rrs N `J#4"5"'"&(-(**,*.!'.2*.#'!%#'#$#$+0MQ&+*--0'*,/-0()*+,1-27A#$JA#4ZA#: A# : A# !4Z A#!4ZA#$JA#"%tA#!4<A#DjA#$;A#DjA#DjA#DjA# "D[!A#"($sN/B*CCdJ#A#$t %A#&!4<'A#($dl)A#* +A#,#T\-A#.$dl/A#0:1A#2!4Z3A#4#Tk5A#6$dl7A#A#r,rN) r6configuration_utilsrmodeling_rope_utilsrutilsr get_loggerr3loggerr__all__r%r,r+rEsA&%31   H %U#'U#p  r,