qiCldZddlmZddlmZddlmZejeZ GddeZ dgZ y)zStableLM model configuration)PreTrainedConfig)RopeParameters)loggingc,<eZdZdZdZdgZ ddedzdedzdedzdedzd edzd edzd edzd edzd edzdedzde dzde dzde e ee fzdzde dzde dzde dzdedzdedzdedzdedzdedzf*fd Z xZ S)StableLmConfiga  This is the configuration class to store the configuration of a [`~StableLmModel`]. It is used to instantiate an StableLM 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 StableLM [stabilityai/stablelm-3b-4e1t](https://huggingface.co/stabilityai/stablelm-3b-4e1t) 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 50304): Vocabulary size of the StableLM model. Defines the number of different tokens that can be represented by the `inputs_ids` passed when calling [`StableLmModel`]. intermediate_size (`int`, *optional*, defaults to 6912): Dimension of the MLP representations. hidden_size (`int`, *optional*, defaults to 2560): Number of hidden layers in the Transformer decoder. 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 encoder. num_key_value_heads (`int`, *optional*, defaults to 32): 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). max_position_embeddings (`int`, *optional*, defaults to 4096): 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). 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-05): The epsilon used by the 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`. tie_word_embeddings (`bool`, *optional*, defaults to `False`): Whether the model's input and output word embeddings should be tied. 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`. use_qkv_bias (`bool`, *optional*, defaults to `False`): Whether or not the model should use bias for qkv layers. qk_layernorm (`bool`, *optional*, defaults to `False`): Whether or not to normalize, per head, the Queries and Keys after projecting the hidden states. use_parallel_residual (`bool`, *optional*, defaults to `False`): Whether to use a "parallel" formulation in each Transformer layer, which can provide a slight training speedup at large scales. hidden_dropout (`float`, *optional*, defaults to 0.0): The dropout ratio after applying the MLP to the hidden states. attention_dropout (`float`, *optional*, defaults to 0.0): The dropout ratio for the attention probabilities. bos_token_id (int, *optional*, defaults to 0): The id of the `BOS` token in the vocabulary. eos_token_id (int, *optional*, defaults to 0): The id of the `EOS` token in the vocabulary. pad_token_id (int, *optional*): The id of the `PAD` token in the vocabulary. Example: ```python >>> from transformers import StableLmModel, StableLmConfig >>> # Initializing a StableLM stablelm-3b style configuration >>> configuration = StableLmConfig() ```stablelmpast_key_valuesN vocab_sizeintermediate_size hidden_sizenum_hidden_layersnum_attention_headsnum_key_value_heads hidden_actmax_position_embeddingsinitializer_rangelayer_norm_eps use_cachetie_word_embeddingsrope_parameters use_qkv_bias qk_layernormuse_parallel_residualhidden_dropoutattention_dropout bos_token_id eos_token_id pad_token_idc n||_||_||_||_||_||_||_||_| |_| |_ | |_ ||_ ||_ ||_ ||_||_| |_|j#dd||_||_||_| |_t-|\di|y)Npartial_rotary_factorg?)r rr r r rrrrrrrrrrrr setdefaultrrrrsuper__init__)selfr r r r rrrrrrrrrrrrrrrrrkwargs __class__s e/opt/pipecat/venv/lib/python3.12/site-packages/transformers/models/stablelm/configuration_stablelm.pyr$zStableLmConfig.__init__is2%'>$&!2!2#6 #6 $!2,"((%:",!2.148(((#6  "6")iii r*r*siluig{Gz?gh㈵>TFNFFFr,r-N)__name__ __module__ __qualname____doc__ model_typekeys_to_ignore_at_inferenceintstrfloatboolrdictr$ __classcell__)r's@r(rrsK ZJ#4"5"'(,"&(**,*,!'.2*.'-!%+0MQ$)$)-2'**-#$#$#'-2#$J2#:2#4Z 2# : 2# !4Z 2#!4Z2#$J2#"%t2#!4<2# 2#$;2#"D[2#($sN/B*CCdJ2#Tk2# Tk!2#" $d{#2#$ %2#&!4<'2#(Dj)2#*Dj+2#,Dj-2#2#r)rN) r1configuration_utilsrmodeling_rope_utilsrutilsr get_loggerr.loggerr__all__r!r)r(r@sA#31   H %C#%C#L  r)