iUddlmZddlmZddlZddlmZddlmZddlm Z m Z ddl m Z dd l mZmZmZdd lmZdd lmZmZmZmZdd lmZmZdd lmZmZddlmZm Z ddl!m"Z"ddl#m$Z$m%Z%m&Z&m'Z'ddl(m)Z)m*Z*ddl+m,Z,e'jZe.Z/edGddej`Z1Gddej`Z2dZ3edde%Gd3d4e=e Z?Gd5d6ee=Z@Gd7d8ee=ZAGd9d:ee=ZBgd;ZCy)>)Callable)OptionalN)nn)ACT2FN)Cache DynamicCache)GenerationMixin)use_kernel_forward_from_hubuse_kernel_func_from_hubuse_kernelized_func)create_causal_mask)GenericForQuestionAnswering GenericForSequenceClassificationGenericForTokenClassificationGradientCheckpointingLayer)BaseModelOutputWithPastCausalLMOutputWithPast)ROPE_INIT_FUNCTIONSdynamic_rope_update)ALL_ATTENTION_FUNCTIONSPreTrainedModel)Unpack)TransformersKwargsauto_docstringcan_return_tuplelogging)check_model_inputsmaybe_autocast) LlamaConfigRMSNormc,eZdZdfd ZdZdZxZS) LlamaRMSNormct|tjt j ||_||_y)z; LlamaRMSNorm is equivalent to T5LayerNorm N)super__init__r Parametertorchonesweightvariance_epsilon)self hidden_sizeeps __class__s r/home/obispo/Crisostomo_bridge/mision_env/lib/python3.12/site-packages/transformers/models/llama/modeling_llama.pyr'zLlamaRMSNorm.__init__5s1 ll5::k#:; #c"|j}|jtj}|j dj dd}|tj ||jzz}|j|j|zS)NT)keepdim) dtypetor)float32powmeanrsqrtr,r+)r- hidden_states input_dtypevariances r1forwardzLlamaRMSNorm.forward=sy#)) %((7  $$Q',,R,>% Ht?T?T4T(UU {{]--k:::r2c^t|jjd|jS)Nz, eps=)tupler+shaper,)r-s r1 extra_reprzLlamaRMSNorm.extra_reprDs* ))*+6$2G2G1HIIr2)gư>)__name__ __module__ __qualname__r'r@rD __classcell__r0s@r1r$r$3s$;Jr2r$c eZdZUejed<d deffd Ze d dedzde dde dzde d e ffd Z ejed ZxZS)LlamaRotaryEmbeddinginv_freqNconfigct||j|_|j|_||_|j j d|_|j}|jdk7rt|j}||j |\}|_ |jd|d|jd|jdy)N rope_typedefaultrLF) persistentoriginal_inv_freq) r&r'max_position_embeddingsmax_seq_len_cachedoriginal_max_seq_lenrMrope_parametersrOcompute_default_rope_parametersrattention_scalingregister_bufferclone)r-rMdevice rope_init_fnrLr0s r1r'zLlamaRotaryEmbedding.__init__Ks "("@"@$*$B$B! 44[A!%!E!E >>Y &.t~~>L+7 V+L($( ZeD 0(..2BuUr2r[z torch.deviceseq_lenreturnz torch.Tensorc|jd}t|ddxs|j|jz}d}d|t j d|dtj j|tj|z zz }||fS) a Computes the inverse frequencies according to the original RoPE implementation Args: config ([`~transformers.PreTrainedConfig`]): The model configuration. device (`torch.device`): The device to use for initialization of the inverse frequencies. seq_len (`int`, *optional*): The current sequence length. Unused for this type of RoPE. Returns: Tuple of (`torch.Tensor`, `float`), containing the inverse frequencies for the RoPE embeddings and the post-processing scaling factor applied to the computed cos/sin (unused in this type of RoPE). rope_thetahead_dimNg?rr4r7)r[r7) rVgetattrr.num_attention_headsr)arangeint64r8float)rMr[r]basedimattention_factorrLs r1rWz4LlamaRotaryEmbedding.compute_default_rope_parameters[s&%%l3fj$/c63E3EIcIc3c U\\!S!5;;?BB&X]XcXcBdgjj k )))r2cN|jddddfjj|jdddj |j }|dddddfj}t |j jtr/|j jdk7r|j jnd}t|d5|j|jzjdd}tj||fd }|j|jz}|j|jz} dddj |j  j |j fS#1swYAxYw) Nrr5r mpscpuF) device_typeenabledr4rirb)rLrgexpandrCr8r[ isinstancetypestrr transposer)catcosrXsinr7) r-x position_idsinv_freq_expandedposition_ids_expandedrnfreqsembrwrxs r1r@zLlamaRotaryEmbedding.forwardysR!MM$4-8>>@GG HZHZ[\H]_acdehhijiqiqr ,QaZ 8 > > @'1!((--'E!((--[`J`ahhmmfk  U C 5&,,.1F1L1L1NNYYZ[]^_E))UEN3C'')d444C'')d444C  5 vvAGGv$cff177f&;;;  5 5s BFF$N)NNN)rErFrGr)Tensor__annotations__r!r' staticmethodrintrBrgrWno_gradrr@rHrIs@r1rKrKHsllV{V %)+/"*d"*(*t* ~u$ % **:U]]_ <