3iadZddlZddlZddlmZddlmZmZddlm Z ddl Z ddl m Z gdZ edZ e d Zed Zed Z d2d e d edede j$dzde f dZ d2d e dedede j$dzde f dZ d2d e deded edede j$dzde fdZd e dede fdZd e de fdZ d2dedeezdzdefdZ d3d e d edede j$dzde f dZ d3d e dedede j$dzde f dZ d4d e deded edede j$dzde fdZd e dede fdZd e de fd Zd e de fd!Zd e de fd"Z d5d e d#ede fd$Z!d e de"eeffd%Z# d6d e d&ede j$dzde fd'Z$ d6d e d&ede j$dzde fd(Z%d e d)edefd*Z& d7d e d ed)edede j$dzde f d+Z' d7d e d ed)edede j$dzde f d,Z( d8d e d&ede j$dzde fd-Z) d9d e d.edede j$dzde f d/Z*d0eee fdeee ffd1Z+e+eZ,e+eZ-e+eZ.e+e Z/e+e!Z0e+e$Z1e+e%Z2e+e'Z3e+e(Z4e+e)Z5e+e*Z6y):zHThis file contains utilities for initializing neural network parameters.N)Callable)LiteralTypeVar) ParamSpec)Tensor)calculate_gainuniform_normal_ trunc_normal_ constant_ones_zeros_eye_dirac_xavier_uniform_xavier_normal_kaiming_uniform_kaiming_normal_ orthogonal_sparse_uniformnormalconstanteyediracxavier_uniform xavier_normalkaiming_uniformkaiming_normal orthogonalsparse_R_P) linearconv1dconv2dconv3dconv_transpose1dconv_transpose2dconv_transpose3dsigmoidtanhrelu leaky_reluselu)fan_infan_outtensorab generatorreturnc~tj5|j|||cdddS#1swYyxYwNr5)torchno_gradr r2r3r4r5s O/var/www/pod-logistic/pod-ai/venv/lib/python3.12/site-packages/torch/nn/init.py_no_grad_uniform_r>Es4 :q!y9:::3<meanstdc~tj5|j|||cdddS#1swYyxYwr8)r:r;r r2r@rAr5s r=_no_grad_normal_rDLs4 >~~dC9~=>>>r?cdtdtfd}||d|zz ks ||d|zzkDrtjddtj5|||z |z }|||z |z }|j d|zdz d|zdz ||j |j|tjd z|j||j|| |cdddS#1swYyxYw) Nxr6cddtj|tjdz zdz S)N?@)matherfsqrt)rFs r=norm_cdfz(_no_grad_trunc_normal_..norm_cdf_s(dhhq499S>122c99zjmean is more than 2 std from [a, b] in nn.init.trunc_normal_. The distribution of values may be incorrect. stacklevelr9rI)minmax) floatwarningswarnr:r;r erfinv_mul_rJrLadd_clamp_) r2r@rAr3r4r5rMlus r=_no_grad_trunc_normal_r^Vs:E:e: q1s7{q1s7{ 2  ;  a$h#% & a$h#% & A 1q519 B   C$))C.() D  ! #+s BC55C>valcxtj5|j|cdddS#1swYyxYwN)r:r;fill_r2r_s r=_no_grad_fill_rds, !||C !!!s09cvtj5|jcdddS#1swYyxYwra)r:r;zero_r2s r=_no_grad_zero_rhs) ||~s/8 nonlinearityparamc\gd}||vs|dk(ry|dk(ry|dk(rtjdS|dk(re|d }nBt|tst|tst|t r|}nt d |d tjdd|d zzz S|d k(r yt d|)aReturn the recommended gain value for the given nonlinearity function. The values are as follows: ================= ==================================================== nonlinearity gain ================= ==================================================== Linear / Identity :math:`1` Conv{1,2,3}D :math:`1` Sigmoid :math:`1` Tanh :math:`\frac{5}{3}` ReLU :math:`\sqrt{2}` Leaky Relu :math:`\sqrt{\frac{2}{1 + \text{negative\_slope}^2}}` SELU :math:`\frac{3}{4}` ================= ==================================================== .. warning:: In order to implement `Self-Normalizing Neural Networks`_ , you should use ``nonlinearity='linear'`` instead of ``nonlinearity='selu'``. This gives the initial weights a variance of ``1 / N``, which is necessary to induce a stable fixed point in the forward pass. In contrast, the default gain for ``SELU`` sacrifices the normalization effect for more stable gradient flow in rectangular layers. Args: nonlinearity: the non-linear function (`nn.functional` name) param: optional parameter for the non-linear function Examples: >>> gain = nn.init.calculate_gain( ... "leaky_relu", 0.2 ... ) # leaky_relu with negative_slope=0.2 .. _Self-Normalizing Neural Networks: https://papers.nips.cc/paper/2017/hash/5d44ee6f2c3f71b73125876103c8f6c4-Abstract.html )r$r%r&r'r(r)r*r+rRr,g?r-rIr.{Gz?znegative_slope z not a valid numberrOr/g?zUnsupported nonlinearity )rJrL isinstanceboolintrU ValueError)rirj linear_fnsnegative_slopes r=rrsLJz!\Y%>    yy~  % =!N5$'5#&%'#Nug5HIJ JyyNA$5 5677    4\NCDDrNctjj|r*tjjt|f||||St ||||S)aFill the input Tensor with values drawn from the uniform distribution. :math:`\mathcal{U}(a, b)`. Args: tensor: an n-dimensional `torch.Tensor` a: the lower bound of the uniform distribution b: the upper bound of the uniform distribution generator: the torch Generator to sample from (default: None) Examples: >>> w = torch.empty(3, 5) >>> nn.init.uniform_(w) r<)r: overrideshas_torch_function_variadichandle_torch_functionr r>r<s r=r r sT( 226:44 vi!qI5   VQ9 55rNctjj|r*tjjt|f||||St ||||S)aFill the input Tensor with values drawn from the normal distribution. :math:`\mathcal{N}(\text{mean}, \text{std}^2)`. Args: tensor: an n-dimensional `torch.Tensor` mean: the mean of the normal distribution std: the standard deviation of the normal distribution generator: the torch Generator to sample from (default: None) Examples: >>> w = torch.empty(3, 5) >>> nn.init.normal_(w) rC)r:rtrurvr rDrCs r=r r sT( 226:44 fYvDcY5   FD#y 99rNc$t||||||S)aFill the input Tensor with values drawn from a truncated normal distribution. The values are effectively drawn from the normal distribution :math:`\mathcal{N}(\text{mean}, \text{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 \text{mean} \leq b`. 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 generator: the torch Generator to sample from (default: None) Examples: >>> w = torch.empty(3, 5) >>> nn.init.trunc_normal_(w) r9)r^)r2r@rAr3r4r5s r=r r s8 "&$QY OOrNctjj|r(tjjt|f||St ||S)zFill the input Tensor with the value :math:`\text{val}`. Args: tensor: an n-dimensional `torch.Tensor` val: the value to fill the tensor with Examples: >>> w = torch.empty(3, 5) >>> nn.init.constant_(w, 0.3) rc)r:rtrurvr rdrcs r=r r +sL 226:44 yS5   &# &&rNct|dS)zFill the input Tensor with the scalar value `1`. Args: tensor: an n-dimensional `torch.Tensor` Examples: >>> w = torch.empty(3, 5) >>> nn.init.ones_(w) rH)rdrgs r=r r =s &# &&rNct|S)zFill the input Tensor with the scalar value `0`. Args: tensor: an n-dimensional `torch.Tensor` Examples: >>> w = torch.empty(3, 5) >>> nn.init.zeros_(w) )rhrgs r=rrJs & !!rNc|jdk7r tdtj5tj|j ||j dddd|S#1swY|SxYw)a=Fill the 2-dimensional input `Tensor` with the identity matrix. Preserves the identity of the inputs in `Linear` layers, where as many inputs are preserved as possible. Args: tensor: a 2-dimensional `torch.Tensor` Examples: >>> w = torch.empty(3, 5) >>> nn.init.eye_(w) rO,Only tensors with 2 dimensions are supported)out requires_gradN) ndimensionrpr:r;rshaperrgs r=rrWsaaGHH Q 6<1, each group of channels preserves identity Args: tensor: a {3, 4, 5}-dimensional `torch.Tensor` groups (int, optional): number of groups in the conv layer (default: 1) Examples: >>> w = torch.empty(3, 16, 5, 5) >>> nn.init.dirac_(w) >>> w = torch.empty(3, 24, 5, 5) >>> nn.init.dirac_(w, 3) )z5Only tensors with 3, 4, or 5 dimensions are supportedrz!dim 0 must be divisible by groupsrRrrOrN)rrpsizerSr:r;rfrange)r2r dimensionssizesout_chans_per_grpmin_dimgds r=rrls ""$J"PQQ KKME Qx&A<==aF*#U1X.G  v A7^ ?PQF10014aQ19LLM1_  --1 A!+ A!+- --1 A!+ A!+ A!+ -  , M-, Ms 1CEEc|j}|dkr td|jd}|jd}d}|jdkDr|jddD]}||z} ||z}||z}||fS)NrOzNFan in and fan out can not be computed for tensor with fewer than 2 dimensionsrRr)dimrprr)r2rnum_input_fmapsnum_output_fmapsreceptive_field_sizesr0r1s r=_calculate_fan_in_and_fan_outrsJA~ \  kk!nO{{1~ zz|aab! &A A %  & 3 3F!55G 7?rNgainct|\}}|tjdt||zz z}tjd|z}t || ||S)aFill the input `Tensor` with values using a Xavier uniform distribution. The method is described in `Understanding the difficulty of training deep feedforward neural networks` - Glorot, X. & Bengio, Y. (2010). The resulting tensor will have values sampled from :math:`\mathcal{U}(-a, a)` where .. math:: a = \text{gain} \times \sqrt{\frac{6}{\text{fan\_in} + \text{fan\_out}}} Also known as Glorot initialization. Args: tensor: an n-dimensional `torch.Tensor` gain: an optional scaling factor generator: the torch Generator to sample from (default: None) Examples: >>> w = torch.empty(3, 5) >>> nn.init.xavier_uniform_(w, gain=nn.init.calculate_gain("relu")) rI@)rrJrLrUr>)r2rr5r0r1rAr3s r=rrsY44F;OFG 3v'7!889 9C #A VaRI 66rNct|\}}|tjdt||zz z}t |d||S)aFill the input `Tensor` with values using a Xavier normal distribution. The method is described in `Understanding the difficulty of training deep feedforward neural networks` - Glorot, X. & Bengio, Y. (2010). The resulting tensor will have values sampled from :math:`\mathcal{N}(0, \text{std}^2)` where .. math:: \text{std} = \text{gain} \times \sqrt{\frac{2}{\text{fan\_in} + \text{fan\_out}}} Also known as Glorot initialization. Args: tensor: an n-dimensional `torch.Tensor` gain: an optional scaling factor generator: the torch Generator to sample from (default: None) Examples: >>> w = torch.empty(3, 5) >>> nn.init.xavier_normal_(w) rI)rrJrLrUrD)r2rr5r0r1rAs r=rrsE24F;OFG 3v'7!889 9C FCi 88rNmodec|j}ddg}||vrtd|d|t|\}}|dk(r|S|S)Nr0r1zMode z" not supported, please use one of )lowerrpr)r2r valid_modesr0r1s r=_calculate_correct_fanrsX ::>> w = torch.empty(3, 5) >>> nn.init.kaiming_uniform_(w, mode="fan_in", nonlinearity="relu") Note: Be aware that ``fan_in`` and ``fan_out`` are calculated assuming that the weight matrix is used in a transposed manner, (i.e., ``x @ w.T`` in ``Linear`` layers, where ``w.shape = [fan_out, fan_in]``). This is important for correct initialization. If you plan to use ``x @ w``, where ``w.shape = [fan_in, fan_out]``, pass in a transposed weight matrix, i.e. ``nn.init.kaiming_uniform_(w.T, ...)``. )r2r3rrir5r,Initializing zero-element tensors is a no-oprOrPrr9N)r:rtrurvrrrVrWrrrJrLr;r ) r2r3rrir5fanrrAbounds r=rrsV 226:44  I%5   FLL DQRS  .C , *D 3 C IIcNS E Cvu BCCCs C--C6c,d|jvrtjdd|St||}t ||}|t j |z }tj5|jd||cdddS#1swYyxYw)aFill the input `Tensor` with values using a Kaiming normal distribution. The method is described in `Delving deep into rectifiers: Surpassing human-level performance on ImageNet classification` - He, K. et al. (2015). The resulting tensor will have values sampled from :math:`\mathcal{N}(0, \text{std}^2)` where .. math:: \text{std} = \frac{\text{gain}}{\sqrt{\text{fan\_mode}}} Also known as He initialization. Args: tensor: an n-dimensional `torch.Tensor` a: the negative slope of the rectifier used after this layer (only used with ``'leaky_relu'``) mode: either ``'fan_in'`` (default) or ``'fan_out'``. Choosing ``'fan_in'`` preserves the magnitude of the variance of the weights in the forward pass. Choosing ``'fan_out'`` preserves the magnitudes in the backwards pass. nonlinearity: the non-linear function (`nn.functional` name), recommended to use only with ``'relu'`` or ``'leaky_relu'`` (default). generator: the torch Generator to sample from (default: None) Examples: >>> w = torch.empty(3, 5) >>> nn.init.kaiming_normal_(w, mode="fan_out", nonlinearity="relu") Note: Be aware that ``fan_in`` and ``fan_out`` are calculated assuming that the weight matrix is used in a transposed manner, (i.e., ``x @ w.T`` in ``Linear`` layers, where ``w.shape = [fan_out, fan_in]``). This is important for correct initialization. If you plan to use ``x @ w``, where ``w.shape = [fan_in, fan_out]``, pass in a transposed weight matrix, i.e. ``nn.init.kaiming_normal_(w.T, ...)``. rrrOrPr9N) rrVrWrrrJrLr:r;r )r2r3rrir5rrrAs r=rrBsV FLL DQRS  .C , *D 3 C ;~~a ~:;;;s ,B  Bc|jdkr td|jdk(r|S|jd}|j|z}|j ||fj dd|}||kr|j tjj|\}}tj|d}|j} || z}||kr|j tj5|j|j||j|ddd|S#1swY|SxYw)aFill the input `Tensor` with a (semi) orthogonal matrix. Described in `Exact solutions to the nonlinear dynamics of learning in deep linear neural networks` - Saxe, A. et al. (2013). The input tensor must have at least 2 dimensions, and for tensors with more than 2 dimensions the trailing dimensions are flattened. Args: tensor: an n-dimensional `torch.Tensor`, where :math:`n \geq 2` gain: optional scaling factor generator: the torch Generator to sample from (default: None) Examples: >>> # xdoctest: +REQUIRES(env:TORCH_DOCTEST_LAPACK) >>> w = torch.empty(3, 5) >>> nn.init.orthogonal_(w) rOz4Only tensors with 2 or more dimensions are supportedrrRr9N)rrpnumelr new_emptyr t_r:linalgqrdiagsignr;view_ascopy_rY) r2rr5rowscols flattenedqrrphs r=rrws,QOPP ||~ ;;q>D <<>T !D  $.66q!y6QI d{  <>> w = torch.empty(3, 5) >>> nn.init.sparse_(w, sparsity=0.1) rOr}rr9N) rrprrJceilr:r;r rrandperm) r2rrAr5rr num_zeroscol_idx row_indices zero_indicess r=rrs.aGHHJD$ (T/*I .q#3T{ .G...K&z 2L,-F<( ) .. M . Ms AB++B5methcjdddtjdtjdtffd }dddd |_|_|S) Nargskwargsr6cZtjdddtd|i|S)Nz `nn.init.z)` is now deprecated in favor of `nn.init.z`.rOrP)rVrW FutureWarning)rrrnew_nameold_names r=deprecated_initz(_make_deprecate..deprecated_inits; z!J8*TV W  T$V$$rNz z_(...) .. warning:: This method is now deprecated in favor of :func:`torch.nn.init.z"`. See :func:`~torch.nn.init.z` for details.)__name__r#rrr"__doc__)rrrrs` @@r=_make_deprecatersy}}H}H%rww%"))%%$ JHIQzR'j :O (O rNra)rrHN)rrHgrIN)rR)rHN)rr0r.N)rRN)rlN)7rrJrVcollections.abcrtypingrrtyping_extensionsrr:r__all__r"r#_NonlinearityType_FanModerU Generatorr>rDr^rdrhrorr r r r r rrrtuplerrrrrrrrrrrrrrrrrrr r!rNr=rsbN $#'  > T]t_    & 'MQ: ::!&:38??T3I: :)- > > > >% >  > )- ) ) ) ) )  ) % ) )X!6!!&! 6f BFGE#GE,/%K$,>GE GEX(, 6 6 6 6% 6  6:(, : : : :% :  ::(, P P P P P  P % P P>'f'5'V'$ '& 'V ' "6 "f "F*26232v2j&U38_.(,7 7 7%7 7F(,9 9 9%9 9>3633c3&2(, >C >C >C >C$ >C % >C  >CF&2(, 2; 2; 2; 2;$ 2; % 2;  2;n(,0 0 0%0 0l(, # ## #% #  #N(2r6*xB/?. ( #  ! 9 %d 1/ !"23 1 [ )  !rN