o ¾e¦ih"ã@sLdZddlmZddlmZddlmZe e¡Z Gdd„deƒZ dgZ dS)zFalcon configurationé)ÚPreTrainedConfig)ÚRopeParameters)Úloggingc2sBeZdZdZdZdgZ              d/dedBdedBdedBdedBdedBdedBdedBdedBdedBdedBdedBdedBdedBdedBdedBd edBd!edBd"e e e e fBdBd#edBd$edBd%edBd&edBd'e dBd(edBf0‡fd)d*„ Z e d+d,„ƒZe d-d.„ƒZ‡ZS)0Ú FalconConfiga´ This is the configuration class to store the configuration of a [`FalconModel`]. It is used to instantiate a Falcon 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 [tiiuae/falcon-7b](https://huggingface.co/tiiuae/falcon-7b) 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 65024): Vocabulary size of the Falcon model. Defines the number of different tokens that can be represented by the `inputs_ids` passed when calling [`FalconModel`] hidden_size (`int`, *optional*, defaults to 4544): Dimension of the hidden representations. num_hidden_layers (`int`, *optional*, defaults to 32): Number of hidden layers in the Transformer decoder. num_attention_heads (`int`, *optional*, defaults to 71): Number of attention heads for each attention layer in the Transformer encoder. num_ln_in_parallel_attn (`int`, *optional*): Set to 2 if separate layer norms are to be used for the MLP and the attention output when using parallel attention, otherwise, 1. layer_norm_epsilon (`float`, *optional*, defaults to 1e-05): The epsilon used by the layer normalization layers. initializer_range (`float`, *optional*, defaults to 0.02): The standard deviation of the truncated_normal_initializer for initializing all weight matrices. use_cache (`bool`, *optional*, defaults to `True`): Whether the model should return the last key/values attentions (not used by all models). Only relevant if `config.is_decoder=True`. hidden_dropout (`float`, *optional*, defaults to 0.0): The dropout probability for MLP layers. attention_dropout (`float`, *optional*, defaults to 0.0): The dropout probability for attention layers. num_kv_heads (`int`, *optional*): Number of key-value heads to use per attention layer. If unset, defaults to the same value as `num_attention_heads`. alibi (`bool`, *optional*, defaults to `False`): Whether to use ALiBi positional biases during self-attention. new_decoder_architecture (`bool`, *optional*, defaults to `False`): Whether to use the new (Falcon-40B) decoder architecture. If `True`, the `multi_query` and `parallel_attn` arguments are ignored, as the new decoder always uses parallel attention. multi_query (`bool`, *optional*, defaults to `True`): Whether to use multi-query attention in the decoder. Ignored when `new_decoder_architecture` is `True`. parallel_attn (`bool`, *optional*, defaults to `True`): Whether to compute attention in parallel with the feedforward layer. If False, they are consecutive instead, as in the original Transformer architecture. Ignored when `new_decoder_architecture` is `True`. bias (`bool`, *optional*, defaults to `False`): Whether to use bias on Linear layers. max_position_embeddings (`int`, *optional*, defaults to 2048): The maximum sequence length that this model might ever be used with, when `alibi` is `False`. Pretrained Falcon models with RoPE support up to 2048 tokens. 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`. bos_token_id (`int`, *optional*, defaults to 11): The id of the "beginning-of-sequence" token. eos_token_id (`int`, *optional*, defaults to 11): The id of the "end-of-sequence" token. pad_token_id (`int`, *optional*): Padding token id. ffn_hidden_size (`int`, *optional*): The hidden size of the feedforward layer in the Transformer decoder. defaults to 4x hidden dim activation (`str`, *optional*, defaults to `"gelu"`): The activation function used in the feedforward layer. tie_word_embeddings (`bool`, *optional*, defaults to `True`): Whether to tie weight embeddings Example: ```python >>> from transformers import FalconModel, FalconConfig >>> # Initializing a small (2-layer) Falcon configuration >>> configuration = FalconConfig(num_hidden_layers=2) >>> # Initializing a model from the small configuration >>> model = FalconModel(configuration) >>> # Accessing the model configuration >>> configuration = model.config ```ÚfalconÚpast_key_valueséþéÀé éGNçñh㈵øä>ç{®Gáz”?TçFéé ÚgeluÚ vocab_sizeÚ hidden_sizeÚnum_hidden_layersÚnum_attention_headsÚnum_ln_in_parallel_attnÚlayer_norm_epsilonÚinitializer_rangeÚ use_cacheÚhidden_dropoutÚattention_dropoutÚ num_kv_headsÚalibiÚnew_decoder_architectureÚ multi_queryÚ parallel_attnÚbiasÚmax_position_embeddingsÚrope_parametersÚ bos_token_idÚ eos_token_idÚ pad_token_idÚffn_hidden_sizeÚ activationÚtie_word_embeddingsc sÞ||_| dd¡}|dur|n||_||_||_||_||_||_| |_| |_ ||_ ||_ ||_ | dur6|n| |_ | |_| |_||_||_||_||_||_||_||_|dur^|d|_n||_||_tƒjdi|¤ŽdS)NÚn_embedé©)rÚpoprrrrrrrrr$r%r&rrrrr r!rr"r(r)r'r#ÚsuperÚ__init__)Úselfrrrrrrrrrrrrrrr r!r"r#r$r%r&r'r(r)Úkwargsr*©Ú __class__r,úm/home/ubuntu/transcripts/venv/lib/python3.10/site-packages/transformers/models/falcon/configuration_falcon.pyr/qs8  zFalconConfig.__init__cCs |j|jS©N)rr©r0r,r,r4Úhead_dim®s zFalconConfig.head_dimcCs|j Sr5)rr6r,r,r4Úrotary²szFalconConfig.rotary)rr r r Nr r TrrNFFTTFrNrrNNrT)Ú__name__Ú __module__Ú __qualname__Ú__doc__Ú model_typeÚkeys_to_ignore_at_inferenceÚintÚfloatÚboolrÚdictÚstrr/Úpropertyr7r8Ú __classcell__r,r,r2r4rs¢Uçþýüûúùø ÷ ö õ ô óòñðïîíìëêéèç= rN) r<Úconfiguration_utilsrÚmodeling_rope_utilsrÚutilsrÚ get_loggerr9ÚloggerrÚ__all__r,r,r,r4Ús