[Th*SSKrSSKrSSKJr SSKJr SSKJr SSKJrJ r SSK J r S/r Sr S rS r"S S\5rg) N)Tensor) constraints) Distribution)_standard_normal lazy_property)_sizeMultivariateNormalcj[R"XRS55RS5$)a Performs a batched matrix-vector product, with compatible but different batch shapes. This function takes as input `bmat`, containing :math:`n \times n` matrices, and `bvec`, containing length :math:`n` vectors. Both `bmat` and `bvec` may have any number of leading dimensions, which correspond to a batch shape. They are not necessarily assumed to have the same batch shape, just ones which can be broadcasted. )torchmatmul unsqueezesqueeze)bmatbvecs _/var/www/auris/envauris/lib/python3.13/site-packages/torch/distributions/multivariate_normal.py _batch_mvrs' <<nnR0 1 9 9" ==cURS5nURSSn[U5nUR5S- nXE- nXe-nUSU--nURSUn [ URSSURUS5HupXU -U 4- n M X4- n UR U 5n[ [U55[ [XhS55-[ [US-US55-U/-n URU 5nUR SX"5n UR SU RS5U5nURSSS5n[RRXSS9RS5RS5nUR5nUR URSS5n[ [U55n[U5HnUUU-UU-/- nM URU5nUR U5$) a7 Computes the squared Mahalanobis distance :math:`\mathbf{x}^\top\mathbf{M}^{-1}\mathbf{x}` for a factored :math:`\mathbf{M} = \mathbf{L}\mathbf{L}^\top`. Accepts batches for both bL and bx. They are not necessarily assumed to have the same batch shape, but `bL` one should be able to broadcasted to `bx` one. r NrFupper)sizeshapelendimzipreshapelistrangepermuter linalgsolve_triangularpowsumt)bLbxnbx_batch_shape bx_batch_dims bL_batch_dimsouter_batch_dimsold_batch_dimsnew_batch_dims bx_new_shapesLsx permute_dimsflat_Lflat_x flat_x_swapM_swapM permuted_Mpermute_inv_dimsi reshaped_Ms r_batch_mahalanobisr?s  AXXcr]N'MFFHqLM$4%5N%M(99N88--.LbhhsmRXX.>r%BCr2& DDL L !B U# $% u%q9 : ; u%)>1= > ?    L !B ZZA !F ZZFKKNA .F..Aq)K %%f%GKKANRRSUV   A288CR=)JE"234 = !-1>A3EFF"##$45J   n --rcr[RR[R"US55n[R"[R"US5SS5n[R "UR SURURS9n[RRX#SS9nU$)N)rr rr dtypedeviceFr) r r$choleskyflip transposeeyerrBrCr%)PLfL_invIdLs r_precision_to_scale_trilrMOs~   uzz!X6 7B OOEJJr84b" =E 1772;aggahh ?B %%eu%=A Hrc^\rSrSrSr\R \R\R\RS.r \R r Sr SU4Sjjr SU4Sjjr \S\4Sj5r\S\4S j5r\S\4S j5r\S\4S j5r\S\4S j5r\S\4S j5r\R0"54S\S\4SjjrSrSrSrU=r$)r Xa) Creates a multivariate normal (also called Gaussian) distribution parameterized by a mean vector and a covariance matrix. The multivariate normal distribution can be parameterized either in terms of a positive definite covariance matrix :math:`\mathbf{\Sigma}` or a positive definite precision matrix :math:`\mathbf{\Sigma}^{-1}` or a lower-triangular matrix :math:`\mathbf{L}` with positive-valued diagonal entries, such that :math:`\mathbf{\Sigma} = \mathbf{L}\mathbf{L}^\top`. This triangular matrix can be obtained via e.g. Cholesky decomposition of the covariance. Example: >>> # xdoctest: +REQUIRES(env:TORCH_DOCTEST_LAPACK) >>> # xdoctest: +IGNORE_WANT("non-deterministic") >>> m = MultivariateNormal(torch.zeros(2), torch.eye(2)) >>> m.sample() # normally distributed with mean=`[0,0]` and covariance_matrix=`I` tensor([-0.2102, -0.5429]) Args: loc (Tensor): mean of the distribution covariance_matrix (Tensor): positive-definite covariance matrix precision_matrix (Tensor): positive-definite precision matrix scale_tril (Tensor): lower-triangular factor of covariance, with positive-valued diagonal Note: Only one of :attr:`covariance_matrix` or :attr:`precision_matrix` or :attr:`scale_tril` can be specified. Using :attr:`scale_tril` will be more efficient: all computations internally are based on :attr:`scale_tril`. If :attr:`covariance_matrix` or :attr:`precision_matrix` is passed instead, it is only used to compute the corresponding lower triangular matrices using a Cholesky decomposition. )loccovariance_matrixprecision_matrix scale_trilTc>UR5S:a [S5eUSLUSL-USL-S:wa [S5eUbjUR5S:a [S5e[R"URSSURSS5nUR US-5UlOUbjUR5S:a [S 5e[R"URSSURSS5nUR US-5UlOiUR5S:a [S 5e[R"URSSURSS5nUR US-5UlUR US -5Ul URRSSn[TU]-XgUS 9 UbX@l gUb%[RRU5Ul g[U5Ul g) Nrz%loc must be at least one-dimensional.zTExactly one of covariance_matrix or precision_matrix or scale_tril may be specified.rzZscale_tril matrix must be at least two-dimensional, with optional leading batch dimensionsrr )r r zZcovariance_matrix must be at least two-dimensional, with optional leading batch dimensionszYprecision_matrix must be at least two-dimensional, with optional leading batch dimensions)r  validate_args)r ValueErrorr broadcast_shapesrexpandrSrQrRrPsuper__init___unbroadcasted_scale_trilr$rDrM) selfrPrQrRrSrV batch_shape event_shape __class__s rr[MultivariateNormal.__init__s 779q=DE E T )j.D E D (  f   !~~!# = 001A1A#21F RUSUWK(// h0FGDO  * $$&* = 00!'',ciinK&7%=%=kH>T%UD "##%) = 00 &&s+SYYs^K%5$;$;K(UR[U5n[R"U5nXR-nXR-UR-nUR R U5UlURUlSUR;a URR U5Ul SUR;a URR U5Ul SUR;a URR U5Ul [[U]7XRSS9 URUlU$)NrQrSrRFrU)_get_checked_instancer r Sizer_rPrYr\__dict__rQrSrRrZr[_validate_args)r]r^ _instancenew loc_shape cov_shaper`s rrYMultivariateNormal.expands(();YGjj- "2"22 "2"22T5E5EE ((//),(,(F(F% $-- /$($:$:$A$A)$LC ! 4== (!__33I>CN  .#'#8#8#?#? #JC   #/ )) 0 "00 rreturncURRURUR-UR-5$N)r\rY _batch_shape _event_shaper]s rrSMultivariateNormal.scale_trils:--44    1 1 1D4E4E E  rc[R"URURR5R UR UR -UR -5$rn)r r r\mTrYrorprqs rrQ$MultivariateNormal.covariance_matrixsQ||  * *D,J,J,M,M &""T%6%669J9JJ K Lrc[R"UR5RURUR -UR -5$rn)r cholesky_inverser\rYrorprqs rrR#MultivariateNormal.precision_matrixsE%%d&D&DELL    1 1 1D4E4E E  rcUR$rnrPrqs rmeanMultivariateNormal.mean xxrcUR$rnrzrqs rmodeMultivariateNormal.moder}rcURRS5RS5RURUR -5$)Nrr )r\r&r'rYrorprqs rvarianceMultivariateNormal.variancesA  * * . .q 1 SW VD%%(9(99 : r sample_shapecURU5n[X RRURRS9nUR[ UR U5-$)NrA)_extended_shaperrPrBrCrr\)r]rrepss rrsampleMultivariateNormal.rsamplesJ$$\2uHHNN488??Sxx)D$B$BCHHHrc|UR(aURU5 XR- n[URU5nURR SSS9R 5RS5nSURS[R "S[R-5-U--U- $)Nrr dim1dim2grr) rf_validate_samplerPr?r\diagonallogr'rpmathpi)r]valuediffr: half_log_dets rlog_probMultivariateNormal.log_probs     ! !% (xx t==t D  * * 3 3" 3 E I I K O OPR S t((+dhhq477{.CCaGH<WWrc\URRSSS9R5RS5nSURS-S[ R"S[ R -5--U-n[UR5S:XaU$URUR5$)Nrr rg?rg?r) r\rrr'rprrrrorY)r]rHs rentropyMultivariateNormal.entropys  * * 3 3" 3 E I I K O OPR S  $##A& &#TWW0E*E F U t  !Q &H88D--. .r)r\rQrPrRrS)NNNNrn)__name__ __module__ __qualname____firstlineno____doc__r real_vectorpositive_definitelower_choleskyarg_constraintssupport has_rsampler[rYrrrSrQrRpropertyr{rrr rdrrrr__static_attributes__ __classcell__)r`s@rr r Xs8"J&&(::'99!00 O %%GK  7Xr& F  L6LL  &  ff &  -2JJLIEIVI X//r)rr rtorch.distributionsr torch.distributions.distributionrtorch.distributions.utilsrr torch.typesr__all__rr?rMr rrrsB +9E  >/.d s/s/r