tj)YdZddlmZddlmZddlZddlmZddlmZddl m Z m Z dd l m Z dd lmZdd lmZdd lmZmZmZdd lmZmZddlmZmZddlmZmZddlm Z ddl!m"Z"m#Z#m$Z$m%Z%ddl&m'Z'm(Z(ddl)m*Z*ddl+m,Z,e%j-e.Z/Gddej0Z1dZ2d4dZ3Gddej0Z4 d5dej0dej5dej5d ej5d!ej5dzd"e6d#e6fd$Z7Gd%d&ej0Z8Gd'd(eZ9e#Gd)d*eZ:e#Gd+d,e:Z;Gd-d.e:e Z<Gd/d0ee:Z=Gd1d2ee:Z>gd3Z?dS)6zPyTorch Persimmon model.)Callable)OptionalN)nn)ACT2FN)Cache DynamicCache)GenerationMixin)create_causal_mask)FlashAttentionKwargs) GenericForSequenceClassificationGenericForTokenClassificationGradientCheckpointingLayer)BaseModelOutputWithPastCausalLMOutputWithPast)ROPE_INIT_FUNCTIONSdynamic_rope_update)ALL_ATTENTION_FUNCTIONSPreTrainedModel)Unpack)TransformersKwargsauto_docstringcan_return_tuplelogging)maybe_autocastmerge_with_config_defaults)capture_outputs)PersimmonConfigc eZdZUejed<d deffd Ze d dedzde dde dzde d e ffd Z ejed ZxZS)PersimmonRotaryEmbeddinginv_freqNconfigct|j|_|j|_||_|jjd|_|j}|jdkrt|j}||j|\}|_ | d|d| d| ddS)N rope_typedefaultr"F) persistentoriginal_inv_freq) super__init__max_position_embeddingsmax_seq_len_cachedoriginal_max_seq_lenr#rope_parametersr%compute_default_rope_parametersrattention_scalingregister_bufferclone)selfr#device rope_init_fnr" __class__s n/var/www/html/CA-Chatbot/venv/lib/python3.11/site-packages/transformers/models/persimmon/modeling_persimmon.pyr*z!PersimmonRotaryEmbedding.__init__<s "("@$*$B! 4[A!%!E >Y & &.t~>L+7< V+L+L($( ZeDDD 0(..2B2BuUUUUUr4z torch.deviceseq_lenreturnz torch.TensorcV|jd}|jdd}t|ddp|j|jz}t ||z}d}d|t jd|dt j |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)r4rA) r.getgetattr hidden_sizenum_attention_headsinttorcharangeint64tofloat) r#r4r9baser=r>dimattention_factorr"s r7r/z8PersimmonRotaryEmbedding.compute_default_rope_parametersLs(%l3 & 6 : :;RTW X X6:t44h8JfNh8h(2233 U\!S!5;???BB&X]XcBddgjj k )))r8cN|jddddf|jddd|j}|dddddf}t |jjtr|jjdkr |jjnd}t|d5||z dd}tj ||fd }| |jz}||jz} dddn #1swxYwY||j | |j fS) NrrmpscpuF) device_typeenabledr?rMr@)r"rKexpandshaperJr4 isinstancetypestrr transposerGcatcosr0sinrA) r3x position_idsinv_freq_expandedposition_ids_expandedrSfreqsembr]r^s r7forwardz PersimmonRotaryEmbedding.forwardms!M$4-8>>@@GG HZ[\H]_acdeehhijiqrr ,QQQaaaZ 8 > > @ @'1!(-'E'Ek!(-[`J`J`ahmmfk  U C C C 5 5&,,..1F1L1L1N1NNYYZ[]^__E)UEN333C''))d44C''))d44C  5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 vvAGv$$cff17f&;&;;;s BE&&E*-E*N)NNN)__name__ __module__ __qualname__rGTensor__annotations__rr* staticmethodrrFtuplerKr/no_gradrre __classcell__r6s@r7r!r!9slVVVVVVVV *.+/"**$&*(*t* ~u$ % ***\*>U]__ < <_ < < < <