o Û¾iZã@s2ddlmZddlmZGdd„deƒZdgZdS)é)ÚAny)ÚPretrainedConfigc2sÒeZdZdZdZdgZ               d/dedededededededededee e fd Bded ed!e d"ed#ed$ed%ed&ed'ed(ed)ed*e d+ed,e e d Bf0‡fd-d.„ Z ‡ZS)0Ú Lfm2MoeConfiga This is the configuration class to store the configuration of a [`Lfm2MoeModel`]. It is used to instantiate a LFM2 Moe 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 LFM2-8B-A1B model. e.g. [LiquidAI/LFM2-8B-A1B](https://huggingface.co/LiquidAI/LFM2-8B-A1B) 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 65536): Vocabulary size of the LLaMA model. Defines the number of different tokens that can be represented by the `inputs_ids` passed when calling [`Lfm2Model`] hidden_size (`int`, *optional*, defaults to 2048): Dimension of the hidden representations. intermediate_size (`int`, *optional*, defaults to 7168): Dimension of the MLP representations. moe_intermediate_size (`int`, *optional*, defaults to 1792): Intermediate size of the routed expert. num_hidden_layers (`int`, *optional*, defaults to 32): Number of hidden layers in the Transformer decoder. pad_token_id (`int`, *optional*, defaults to 0): 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 `True`): Whether to tie weight embeddings rope_parameters (`dict`, *optional*): The parameters of the RoPE embeddings. max_position_embeddings (`int`, *optional*, defaults to 128000): The maximum sequence length that this model might ever be used with. 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`. norm_eps (`float`, *optional*, defaults to 1e-05): The epsilon used by the rms normalization layers. 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*, defaults to 8): 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`. conv_bias (`bool`, *optional*, defaults to `False`): Whether to use bias in the conv layers. conv_L_cache (`int`, *optional*, defaults to 3): L_cache dim in the conv layers. num_dense_layers (`int`, *optional*, defaults to 2): Number of dense Lfm2MoeMLP layers in shallow layers(embed->dense->dense->...->dense->moe->moe...->lm_head). num_experts_per_tok (`int`, *optional*, defaults to 4): Number of selected experts. num_experts (`int`, *optional*, defaults to 32): Number of routed experts. use_expert_bias (`bool`, *optional*, defaults to `True`): Whether to use the expert bias on the routing weights. routed_scaling_factor (`float`, *optional*, defaults to 1.0): Scaling factor for routed experts in MoE models. norm_topk_prob (`bool`, *optional*, defaults to `True`): Whether to normalize the topk probabilities. layer_types (`Optional`, *optional*): Type of each layers. ```python >>> from transformers import Lfm2MoeModel, Lfm2MoeConfig >>> # Initializing a LFM2 Moe model >>> configuration = Lfm2MoeConfig() >>> # Initializing a model from the LFM2-8B-A1B style configuration >>> model = Lfm2MoeModel(configuration) >>> # Accessing the model configuration >>> configuration = model.config ```Úlfm2_moeÚpast_key_valuesééééé rééTNéôçñh㈵øä>éFééçð?Ú vocab_sizeÚ hidden_sizeÚintermediate_sizeÚmoe_intermediate_sizeÚnum_hidden_layersÚ pad_token_idÚ bos_token_idÚ eos_token_idÚtie_word_embeddingsÚrope_parametersÚmax_position_embeddingsÚ use_cacheÚnorm_epsÚnum_attention_headsÚnum_key_value_headsÚ conv_biasÚ conv_L_cacheÚnum_dense_layersÚnum_experts_per_tokÚ num_expertsÚuse_expert_biasÚrouted_scaling_factorÚnorm_topk_probÚ layer_typesc sÂ||_||_||_||_| dd¡}| durd|dœ} | |_| |_| |_| |_||_ ||_ ||_ ||_ ||_ ||_||_||_||_||_||_||_| d| ¡} tƒjd|||| dœ|¤ŽdS)NÚ rope_thetag€„.AÚdefault)Ú rope_typer,Ú tie_embedding)rrrr©)rrrrÚpoprrrr r!r"r#r$r%rr&r'r(r)r*r+ÚgetÚsuperÚ__init__)Úselfrrrrrrrrrrrrr r!r"r#r$r%r&r'r(r)r*r+Úkwargsr,©Ú __class__r0ú\/home/ubuntu/.local/lib/python3.10/site-packages/vllm/transformers_utils/configs/lfm2_moe.pyr4]sD  ÿü ûzLfm2MoeConfig.__init__)rrr r r rr r TNrTrr rFrr rr TrTN)Ú__name__Ú __module__Ú __qualname__Ú__doc__Ú model_typeÚkeys_to_ignore_at_inferenceÚintÚboolÚdictÚstrrÚfloatÚlistr4Ú __classcell__r0r0r7r9rsšQçþýüûúùø ÷ ö õ ô óòñðïîíìëêéè çrN)ÚtypingrÚ transformers.configuration_utilsrrÚ__all__r0r0r0r9Ús