`i]dZddlZddlZddlmZddlmZmZmZm Z ddl Z ddl Z ddl m Z ddlmZddlmZdd lmZdd lmZdd lmZmZmZdd lmZmZdd lmZddlm Z m!Z!m"Z"m#Z#m$Z$m%Z%ddl&m'Z'ddl(m)Z)m*Z*m+Z+dZ, dTde j-de.de.de.de.de j-f dZ/dUd Z0d!Z1d"Z2ee"d#$Gd%d&e Z3ee"d'$Gd(d)e Z4ee"Gd*d+e Z5Gd,d-e j6Z7Gd.d/e j6Z8 dVd0e j6d1e j-d2e j-d3e j-d4ee j-d5e.d6e.fd7Z9Gd8d9e j6Z:Gd:d;e j6Z;Gd<d=eZd?eZ=Gd@dAe j6Z>GdBdCe j6Z?e"dD$GdEdFe=Z@GdGdHe j6ZAGdIdJe j6ZBe"dK$GdLdMe=ZCe"GdNdOe=ZDe"dP$GdQdRe=ZEgdSZFdS)WzPyTorch Siglip model.N) dataclass)AnyCallableOptionalUnion)nn)_calculate_fan_in_and_fan_out)ACT2FN)_prepare_4d_attention_mask)GradientCheckpointingLayer)BaseModelOutputBaseModelOutputWithPoolingImageClassifierOutput)ALL_ATTENTION_FUNCTIONSPreTrainedModel)Unpack) ModelOutputTransformersKwargsauto_docstringcan_return_tuplefilter_out_non_signature_kwargs torch_int)check_model_inputs) SiglipConfigSiglipTextConfigSiglipVisionConfigcd}||d|zz ks ||d|zzkrtjdd|||z |z }|||z |z }|d|zdz d|zdz |||t jdz|||||dS)Nc`dtj|tjdz zdz S)N?@)matherfsqrt)xs ~/home/jaya/work/projects/VOICE-AGENT/VIET/agent-env/lib/python3.11/site-packages/transformers/models/siglip/modeling_siglip.pynorm_cdfz _trunc_normal_..norm_cdf0s)dhq49S>>1222c99zjmean is more than 2 std from [a, b] in nn.init.trunc_normal_. The distribution of values may be incorrect.) stacklevelrr")minmax) warningswarnuniform_erfinv_mul_r#r%add_clamp_)tensormeanstdabr(lus r'_trunc_normal_r<-s::: q1s7{q1s7{ 2 2  ;     !d(c!""A!d(c!""A OOAEAIq1uqy))) NN KKdinn$%%% KK MMaQMr)r!r"r5r6r7r8r9returnctj5t|dd|||||ddddS#1swxYwYdS)anFills the input Tensor with values drawn from a truncated normal distribution. The values are effectively drawn from the normal distribution :math:`\mathcal{N}( ext{mean}, ext{std}^2)` with values outside :math:`[a, b]` redrawn until they are within the bounds. The method used for generating the random values works best when :math:`a \leq ext{mean} \leq b`. NOTE: this 'tf' variant behaves closer to Tensorflow / JAX impl where the bounds [a, b] are applied when sampling the normal distribution with mean=0, std=1.0 and the result is subsequently scaled and shifted by the mean and std args. Args: tensor: an n-dimensional `torch.Tensor` mean: the mean of the normal distribution std: the standard deviation of the normal distribution a: the minimum cutoff value b: the maximum cutoff value rr!N)torchno_gradr<r2r3)r5r6r7r8r9s r'trunc_normal_tf_rCQs* $$vq#q!,,, Cd###$$$$$$$$$$$$$$$$$$s