`iq?jddlmZmZmZddlmZmZGddeZGddeZddgZ dS) )AnyOptionalUnion)PretrainedConfiglayer_type_validationceZdZdZdZdgZddddddddZdgdgfd d gd gfd gd gfd Z d#fd" ZxZ S)$T5GemmaModuleConfigaH This is the configuration class to store the configuration of a [`T5GemmaModuleModel`]. It is used to instantiate an T5GemmaModule 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 T5GemmaModule-7B. e.g. [google/t5_gemma_module-7b](https://huggingface.co/google/t5_gemma_module-7b) 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 256000): Vocabulary size of the T5GemmaModule model. Defines the number of different tokens that can be represented by the `inputs_ids` passed when calling [`T5GemmaModuleModel`] hidden_size (`int`, *optional*, defaults to 2304): Dimension of the hidden representations. intermediate_size (`int`, *optional*, defaults to 9216): Dimension of the MLP representations. num_hidden_layers (`int`, *optional*, defaults to 26): Number of hidden layers in the Transformer decoder. num_attention_heads (`int`, *optional*, defaults to 8): Number of attention heads for each attention layer in the Transformer decoder. num_key_value_heads (`int`, *optional*, defaults to 4): 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`. head_dim (`int`, *optional*, defaults to 256): The attention head dimension. hidden_activation (`str` or `function`, *optional*, defaults to `"gelu_pytorch_tanh"`): The non-linear activation function (function or string) in the decoder. Will default to `"gelu_pytorch_tanh"` if not specified. `"gelu_pytorch_tanh"` uses an approximation of the `"gelu"` activation function. max_position_embeddings (`int`, *optional*, defaults to 8192): 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*, defaults to 0): Padding token id. eos_token_id (`int`, *optional*, defaults to 1): End of stream token id. bos_token_id (`int`, *optional*, defaults to 2): Beginning of stream token id. tie_word_embeddings (`bool`, *optional*, defaults to `True`): Whether to tie weight embeddings rope_theta (`float`, *optional*, defaults to 10000.0): The base period of the RoPE embeddings. attention_bias (`bool`, defaults to `False`, *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. query_pre_attn_scalar (`float`, *optional*, defaults to 256): scaling factor used on the attention scores sliding_window (`int`, *optional*, defaults to 4096): in T5GemmaModule, every other layer uses sliding window attention. This is the size of the sliding window. layer_types (`list`, *optional*): Attention pattern for each layer. final_logit_softcapping (`float`, *optional*, defaults to 30.0): scaling factor when applying tanh softcapping on the logits. attn_logit_softcapping (`float`, *optional*, defaults to 50.0): scaling factor when applying tanh softcapping on the attention scores. ```python >>> from transformers import T5GemmaModuleModel, T5GemmaModuleConfig >>> # Initializing a T5GemmaModule t5_gemma_module-7b style configuration >>> configuration = T5GemmaModuleConfig() >>> # Initializing a model from the t5_gemma_module-7b style configuration >>> model = T5GemmaModuleModel(configuration) >>> # Accessing the model configuration >>> configuration = model.config ```t5_gemma_modulepast_key_valuescolwiserowwise)zlayers.*.self_attn.q_projzlayers.*.self_attn.k_projzlayers.*.self_attn.v_projzlayers.*.self_attn.o_projzlayers.*.mlp.gate_projzlayers.*.mlp.up_projzlayers.*.mlp.down_proj input_ids inputs_embeds hidden_statesattention_mask) embed_tokenslayersnorm $gelu_pytorch_tanh {Gz?ư>Tr@FN>@I@c tjd| |||d|||_| |_||_||_||_||_||_||_ | |_ | |_ | |_ ||_ ||_||_||_||_||_||_||_||_|j#dt-|jD|_t/|j|jdS)N) pad_token_id bos_token_id eos_token_idtie_word_embeddingsc@g|]}t|dzdzrdndS)r!r"sliding_attentionfull_attention)bool).0is /home/jaya/work/projects/VOICE-AGENT/VIET/agent-env/lib/python3.11/site-packages/transformers/models/t5gemma/configuration_t5gemma.py z0T5GemmaModuleConfig.__init__..sA   STtQUaK'8'8N##>N   )super__init__ vocab_sizemax_position_embeddings hidden_sizeintermediate_sizenum_hidden_layersnum_attention_headshead_dimnum_key_value_headsinitializer_range rms_norm_eps use_cache rope_thetaattention_biasattention_dropouthidden_activationquery_pre_attn_scalarsliding_windowfinal_logit_softcappingattn_logit_softcapping layer_typesranger)selfr9r;r<r=r>r@r?rGr:rArBrCr)r+r*r,rDrErFrHrIrLrJrKkwargs __class__s r3r8zT5GemmaModuleConfig.__init__ys28  %%% 3       %'>$&!2!2#6   #6 !2("$,!2!2%:",'>$&<#&   #  X]^b^tXuXu   D  d.0FGGGGGr5)rrrrrrrrrrr Trr!r"Tr#Fr$rr%Nr&r') __name__ __module__ __qualname____doc__ model_typekeys_to_ignore_at_inferencebase_model_tp_planbase_model_pp_planr8 __classcell__rPs@r3r r s J J X#J#4"5%.%.%.%."+ )"+&(9:#%568IJ!"_$56- $ ! $#3<H<H<H<H<H<H<H<H<H<Hr5r czeZdZdZdZdgZidddddddd d dd dd d d ddddddd dddddddd dddddd iZdgdgfddgdgfdgdgfdgdgfddgdgfdgdgfdZ d+d!ee e e e e ffd"ee e e e e ffd#e d$ed%ed&ed'e d(effd) Zfd*ZxZS), T5GemmaConfiga This is the configuration class to store the configuration of a [`T5GemmaModel`]. It is used to instantiate an T5Gemma model according to the specified arguments, defining the model architecture. Instantiating a configuration with the defaults will yield a similar configuration to a hypothetical balanced Gemma2 encoder-decoder model. e.g. [google/t5gemma-2b-2b-prefixlm-it](https://huggingface.co/google/t5gemma-2b-2b-prefixlm-it) ```python >>> from transformers import T5GemmaConfig, T5GemmaModel >>> t5gemma_config = T5GemmaConfig.from_pretrained("google/t5gemma-2b-2b-prefixlm-it") >>> model = T5GemmaModel(t5gemma_config) ``` Configuration objects inherit from [PretrainedConfig] and can be used to control the model outputs. Read the documentation from [PretrainedConfig] for more information. Args: encoder (`Union[T5GemmaModuleConfig, dict]`, optional, *optional*): Configuration for the encoder. decoder (`Union[T5GemmaModuleConfig, dict]`, optional, *optional*): Configuration for the decoder. is_encoder_decoder (bool, optional, *optional*, defaults to `True`): Whether the model is used as an encoder/decoder or not. dropout_rate (`float`, *optional*, defaults to 0.0): The ratio for all dropout layers (following T5). classifier_dropout_rate (`float`, *optional*, defaults to 0.0): The dropout ratio for classifier (following T5). attention_dropout (`float`, *optional*, defaults to 0.0): The dropout ratio for attention. tie_word_embeddings (`bool`, *optional*, defaults to `True`): Whether tie input and output embeddings. vocab_size (`int`, *optional*, defaults to 256000): Vocabulary size of the T5Gemma model (the same as Gemma 2). kwargs (additional keyword arguments, optional, *optional*): Will be passed to the PretrainedConfig base class. t5gemmar z!encoder.layers.*.self_attn.q_projr z!encoder.layers.*.self_attn.k_projz!encoder.layers.*.self_attn.v_projz!encoder.layers.*.self_attn.o_projrzencoder.layers.*.mlp.gate_projzencoder.layers.*.mlp.up_projzencoder.layers.*.mlp.down_projz!decoder.layers.*.self_attn.q_projz!decoder.layers.*.self_attn.k_projz!decoder.layers.*.self_attn.v_projz!decoder.layers.*.self_attn.o_projz"decoder.layers.*.cross_attn.q_projz"decoder.layers.*.cross_attn.k_projz"decoder.layers.*.cross_attn.v_projz"decoder.layers.*.cross_attn.o_projzdecoder.layers.*.mlp.gate_projzdecoder.layers.*.mlp.up_projzdecoder.layers.*.mlp.down_projrrrr)zencoder.embed_tokenszencoder.layersz encoder.normzdecoder.embed_tokenszdecoder.layersz decoder.normNTr$rencoderdecoderis_encoder_decoder dropout_rateclassifier_dropout_raterFr,r9c  t|tr tdi|}n@|t}n/t|tsJt|dt|tr tdi|}n4||}n/t|tsJt|dtdi|}tdi|}d|_||_||_||_d|_d|_ ||_||_|j |_ ||_ dD]} | | vrt|| | | <tjdi| ||_| d|j |_ | d|j|_||_||_||_||_||_dS)Nz is not supported.FT)r*r)r+rCrAr6) isinstancedictr typeto_dict is_decoderrarFr^rCr;cross_attention_hidden_sizer_getattrr7r8r`getrArbr,r9) rNr^r_r`rarbrFr,r9rOspecial_token_keyrPs r3r8zT5GemmaConfig.__init__s gt $ $ b)44G44GG _)++GGg':;; a aW =a=a=a a a; gt $ $ b)44G44GG _GGg':;; a aW =a=a=a a a;%::(9(9::%::(9(9::"+$5! ! +$5!.5.A+ !Q P P  ..,3G=N,O,O()""6""""4K1BCC!',?AZ![![(!2'>$#6 %r5cgd}||vr,t|j||t|j||t||dS)N)output_hidden_statesoutput_attentions_attn_implementationrarFr9)setattrr^r_r7 __setattr__)rNkeyvalueshared_attr_with_submodulesrPs r3rrzT5GemmaConfig.__setattr__7sk' ' ' # - - - DL#u - - - DL#u - - - C'''''r5)NNTr$r$r$Tr)rQrRrSrTrUrVrWrXrrr rerr0floatintr8rrrYrZs@r3r\r\s[BJ#4"5+Y ,Y ,Y  ,Y  ))  '  )) ,Y ,Y ,Y ,Y -i -i -i!" -i#$ ))%& ' '( )))0#.0A B+-=>@QR)*_,=>"-0A B+-=>@QR)*_,=>  IMHL#'!),#&$( 8%8%% 3T#s(^ CDE8%% 3T#s(^ CDE8%! 8%  8% "' 8%!8%"8%8%8%8%8%8%8%t ( ( ( ( ( ( ( ( (r5r\N) typingrrrconfiguration_utilsrrr r\__all__r6r5r3r{s,('''''''''JJJJJJJJZHZHZHZHZH*ZHZHZHzL(L(L(L(L($L(L(L(^ 1 2r5