^jqQ@ddlmZddlmZddlZddlmZddlmZddl m Z m Z ddl m Z ddlmZmZdd lmZdd lmZmZmZdd lmZmZdd lmZmZdd lmZmZddl m!Z!ddl"m#Z#m$Z$ddl%m&Z&m'Z'm(Z(ddl)m*Z*ddl+m,Z,GddejZZ.dZ/edd7dZ0dejbde2dejbfdZ3 d8dejZdejbdejbd ejbd!ejbdzd"e4d#e4d$e!e#fd%Z5ee0Gd&d'ejZZ6Gd(d)ejZZ7Gd*d+eZ8e$Gd,d-eZ9e$Gd.d/e9Z:e$Gd0d1e9e Z;Gd2d3ee9Z<Gd4d5ee9Z=gd6Z>y)9)Callable)OptionalN)ACT2FN)Cache DynamicCache)GenerationMixin)use_kernel_func_from_hubuse_kernelized_func)create_causal_mask) GenericForSequenceClassificationGenericForTokenClassificationGradientCheckpointingLayer)BaseModelOutputWithPastCausalLMOutputWithPast)ROPE_INIT_FUNCTIONSdynamic_rope_update)ALL_ATTENTION_FUNCTIONSPreTrainedModel)Unpack)TransformersKwargsauto_docstring)can_return_tuplemaybe_autocastmerge_with_config_defaults)capture_outputs) PhiConfigc eZdZUejed<d deffd Ze d dedzde dde dzde d e ffd Z ejed ZxZS)PhiRotaryEmbeddinginv_freqNconfigct||j|_|j|_||_|j j d|_|j}|jdk7rt|j}||j |\}|_ |jd|d|jd|jdy)N rope_typedefaultr!F) persistentoriginal_inv_freq) super__init__max_position_embeddingsmax_seq_len_cachedoriginal_max_seq_lenr"rope_parametersr$compute_default_rope_parametersrattention_scalingregister_bufferclone)selfr"device rope_init_fnr! __class__s o/var/www/ramen.bs-engineer-server.com/venv/lib/python3.12/site-packages/transformers/models/phi/modeling_phi.pyr)zPhiRotaryEmbedding.__init__$s "("@"@$*$B$B! 44[A!%!E!E >>Y &.t~~>L+7 V+L($( ZeD 0(..2BuUr3z torch.deviceseq_lenreturnz torch.Tensorcn|jd}|jjdd}t|ddxs|j|jz}t ||z}d}d|t jd|dt jj|t j |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_thetapartial_rotary_factorg?head_dimNrdtype)r3r@) r-getgetattr hidden_sizenum_attention_headsinttorcharangeint64tofloat) r"r3r8baser<r=dimattention_factorr!s r6r.z2PhiRotaryEmbedding.compute_default_rope_parameters4s&%%l3 & 6 6 : :;RTW X6:t4h8J8JfNhNh8h(223 U\\!S!5;;?BB&X]XcXcBdgjj k )))r7cN|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) NrrmpscpuF) device_typeenabledr>rLr?)r!rJexpandshaperIr3 isinstancetypestrr transposerFcatcosr/sinr@) r2x position_idsinv_freq_expandedposition_ids_expandedrRfreqsembr\r]s r6forwardzPhiRotaryEmbedding.forwardTsR!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$N)NNN)__name__ __module__ __qualname__rFTensor__annotations__rr) staticmethodrrEtuplerJr.no_gradrrd __classcell__r5s@r6r r !sllVyV #'+/"*D *(*t* ~u$ % **>U]]_ <