+ j5^RIt^RIHt^RIHtHtHt^RIt^RIHt^RI H t H t ]PPPt!RR]PP4t!RR ]PP"4t!R R ]4t!R R ]4t!RR]PP4t!RR]PP"4t!RR]PP4tR#)N)Sequence)CallableOptionalUnionTensor)_log_api_usage_once _make_ntuplecaa]tRt^toRtR V3RlV3RllltV3RlV3RlltV3RlRltV3RlR ltR t Vt V;t #) FrozenBatchNorm2da  BatchNorm2d where the batch statistics and the affine parameters are fixed Args: num_features (int): Number of features ``C`` from an expected input of size ``(N, C, H, W)`` eps (float): a value added to the denominator for numerical stability. Default: 1e-5 c&<V^8dQhRS[RS[/#) num_featureseps)intfloat)format __classdict__s"U/home/luca/screen-checking/.venv/lib/python3.14/site-packages/torchvision/ops/misc.py __annotate__FrozenBatchNorm2d.__annotate__s# F F F Fcv<\SV`4\V4W nVP R\ P !V44VP R\ P!V44VP R\ P!V44VP R\ P !V44R#)weightbias running_mean running_varN)super__init__r rregister_buffertorchoneszeros)selfrr __class__s&&&rrFrozenBatchNorm2d.__init__s D! Xuzz,'?@ VU[[%>? ^U[[-FG ]EJJ|,DErct<V^8dQhRS[RS[RS[RS[RS[S[,RS[S[,RS[S[,/#)r state_dictprefixlocal_metadatastrict missing_keysunexpected_keys error_msgs)dictstrboollist)rrs"rrr$s]        3i  c I rc N<VR,pW9dW\S V`WW4WVV4R#)num_batches_trackedN)r_load_from_state_dict) r$r(r)r*r+r,r-r.num_batches_tracked_keyr%s &&&&&&&& rr5'FrozenBatchNorm2d._load_from_state_dict$s4#)+@"@ " 03 %  Wa rc&<V^8dQhRS[RS[/#rxreturnr)rrs"rrr6s     F  rc|VPP^R^^4pVPP^R^^4pVPP^R^^4pVPP^R^^4pW$VP ,P 4,pW5V,, pW,V,#))rreshaperrrrrsqrt)r$r:wbrvrmscalers&& rforwardFrozenBatchNorm2d.forward6s KK  2q! , II  aQ *    % %aQ 2    & &q"a 3$((]))++:~y4rc <V^8dQhRS[/#)rr;)r0)rrs"rrrAsTT#TrcVPP RVPP^, RVP R2#)(z, eps=))r%__name__rshaper)r$s&r__repr__FrozenBatchNorm2d.__repr__As;..))*!DKK,=,=a,@+AzQRSSr)r)gh㈵>) rL __module__ __qualname____firstlineno____doc__rr5rFrN__static_attributes____classdictcell__ __classcell__r%rs@@rr r s< F F  $   TTTrr c aa]tRt^Eto^^R^]P P ]P P^RR]P P3 V3RlV3Rlllt Rt Vt V;t #)ConvNormActivationNTc<V^8dQhRS[RS[RS[S[S[S[R3,3,RS[S[S[S[R3,3,RS[S[S[S[S[R3,S[3,,RS[RS[S[RS[PP3,,R S[S[RS[PP3,,R S[S[S[S[R3,3,R S[S[ ,R S[S[ ,R S[RS[PP3,RR/ #)r in_channels out_channels kernel_size.stridepaddinggroups norm_layeractivation_layerdilationinplacer conv_layerr;N rrtuplerr0rr!nnModuler1)rrs"rrConvNormActivation.__annotate__Fs$55553c3h/0 5 c5c?*+ 5 %U38_c 9:; 55Xc588??&:;<5#8C,@#AB5U38_,-5$5tn5S%((//125 5rc <aa Vf\S\4'd/\S \4'dS^, ^,S ,pM\S\4'd \S4M \S 4p \ SV 4o\ S V 4o \ ;QJd!.V V3Rl\ V 44FNK 5M!V V3Rl\ V 444pV fVRJp V !VVSVVS VV R7.pVeVPV!V44Ve"V f/MRV /pVPV!R/VB4\SV`$!V!\V4W n VP\8Xd\P!R4R#R#)Nc3n<"TF*pSV,^, ^,SV,,xK, R#5ir=N).0ircr]s& r .ConvNormActivation.__init__..]s-bQaAQ!!3 9HQK G GQas25)rcr`rrdzhDon't use ConvNormActivation directly, please use Conv2dNormActivation and Conv3dNormActivation instead.rn) isinstancerrlenr rgrangeappendrrr r\r%rYwarningswarn)r$r[r\r]r^r_r`rarbrcrdrre _conv_dimlayersparamsr%s&&&f&&&&&f&&& rrConvNormActivation.__init__FsQ ?+s++ 8S0I0I&?q08;0:;0Q0QC ,WZ[cWd *; B ')<%bQVW`Qab%%bQVW`Qabb <%D !     ! MM*\2 3  '"?RG0DF MM*4V4 5 &!D!( >>/ / MMz  0r)r\) rLrPrQrRr!rh BatchNorm2dReLUConv2drrTrUrVrWs@@rrYrYEsN 45./>B?Dxx?S?SEJXX]]01"&#5:XX__555rrYc aa]tRt^~toRt^^R^]P P]P P^RR3 V3RlV3Rlllt Rt Vt V;t #)Conv2dNormActivationa Configurable block used for Convolution2d-Normalization-Activation blocks. Args: in_channels (int): Number of channels in the input image out_channels (int): Number of channels produced by the Convolution-Normalization-Activation block kernel_size: (int, optional): Size of the convolving kernel. Default: 3 stride (int, optional): Stride of the convolution. Default: 1 padding (int, tuple or str, optional): Padding added to all four sides of the input. Default: None, in which case it will be calculated as ``padding = (kernel_size - 1) // 2 * dilation`` groups (int, optional): Number of blocked connections from input channels to output channels. Default: 1 norm_layer (Callable[..., torch.nn.Module], optional): Norm layer that will be stacked on top of the convolution layer. If ``None`` this layer won't be used. Default: ``torch.nn.BatchNorm2d`` activation_layer (Callable[..., torch.nn.Module], optional): Activation function which will be stacked on top of the normalization layer (if not None), otherwise on top of the conv layer. If ``None`` this layer won't be used. Default: ``torch.nn.ReLU`` dilation (int): Spacing between kernel elements. Default: 1 inplace (bool): Parameter for the activation layer, which can optionally do the operation in-place. Default ``True`` bias (bool, optional): Whether to use bias in the convolution layer. By default, biases are included if ``norm_layer is None``. NTc<V^8dQhRS[RS[RS[S[S[S[S[3,3,RS[S[S[S[S[3,3,RS[S[S[S[S[S[3,S[3,,RS[RS[S[RS[PP3,,R S[S[RS[PP3,,R S[S[S[S[S[3,3,R S[S[ ,R S[S[ ,R R/ #rr[r\r]r^r_r`ra.rbrcrdrr;Nrf)rrs"rr!Conv2dNormActivation.__annotate__s    3c3h/0  c5c?*+  %U38_c 9:;   Xc588??&:;< #8C,@#AB U38_,- $ tn   rc l<\S V`VVVVVVVVV V V \PP4 R#N)rrr!rhr r$r[r\r]r^r_r`rarbrcrdrr%s &&&&&&&&&&&&rrConv2dNormActivation.__init__>             HHOO rrn) rLrPrQrRrSr!rhr}r~rrTrUrVrWs@@rrr~sH,45./>B?Dxx?S?SEJXX]]01"&#   rrc aa]tRt^toRt^^R^]P P]P P^RR3 V3RlV3Rlllt Rt Vt V;t #)Conv3dNormActivationa Configurable block used for Convolution3d-Normalization-Activation blocks. Args: in_channels (int): Number of channels in the input video. out_channels (int): Number of channels produced by the Convolution-Normalization-Activation block kernel_size: (int, optional): Size of the convolving kernel. Default: 3 stride (int, optional): Stride of the convolution. Default: 1 padding (int, tuple or str, optional): Padding added to all four sides of the input. Default: None, in which case it will be calculated as ``padding = (kernel_size - 1) // 2 * dilation`` groups (int, optional): Number of blocked connections from input channels to output channels. Default: 1 norm_layer (Callable[..., torch.nn.Module], optional): Norm layer that will be stacked on top of the convolution layer. If ``None`` this layer won't be used. Default: ``torch.nn.BatchNorm3d`` activation_layer (Callable[..., torch.nn.Module], optional): Activation function which will be stacked on top of the normalization layer (if not None), otherwise on top of the conv layer. If ``None`` this layer won't be used. Default: ``torch.nn.ReLU`` dilation (int): Spacing between kernel elements. Default: 1 inplace (bool): Parameter for the activation layer, which can optionally do the operation in-place. Default ``True`` bias (bool, optional): Whether to use bias in the convolution layer. By default, biases are included if ``norm_layer is None``. NTc<V^8dQhRS[RS[RS[S[S[S[S[S[3,3,RS[S[S[S[S[S[3,3,RS[S[S[S[S[S[S[3,S[3,,RS[RS[S[RS[PP3,,R S[S[RS[PP3,,R S[S[S[S[S[S[3,3,R S[S[ ,R S[S[ ,R R/ #rrf)rrs"rr!Conv3dNormActivation.__annotate__s    3c3m 445  c5c3//0  %U3S=%93 >?@   Xc588??&:;< #8C,@#AB U3S=112 $ tn   rc l<\S V`VVVVVVVVV V V \PP4 R#r)rrr!rhConv3drs &&&&&&&&&&&&rrConv3dNormActivation.__init__rrrn) rLrPrQrRrSr!rh BatchNorm3dr~rrTrUrVrWs@@rrrsH*9:34CG?Dxx?S?SEJXX]]56"&#   rrcaa]tRt^toRt]P P]P P3V3RlV3Rlllt V3RlRlt V3RlRlt Rt Vt V;t#) SqueezeExcitationa% This block implements the Squeeze-and-Excitation block from https://arxiv.org/abs/1709.01507 (see Fig. 1). Parameters ``activation``, and ``scale_activation`` correspond to ``delta`` and ``sigma`` in eq. 3. Args: input_channels (int): Number of channels in the input image squeeze_channels (int): Number of squeeze channels activation (Callable[..., torch.nn.Module], optional): ``delta`` activation. Default: ``torch.nn.ReLU`` scale_activation (Callable[..., torch.nn.Module]): ``sigma`` activation. Default: ``torch.nn.Sigmoid`` c <V^8dQhRS[RS[RS[RS[PP3,RS[RS[PP3,RR/#)rinput_channelssqueeze_channels activation.scale_activationr;N)rrr!rhri)rrs"rrSqueezeExcitation.__annotate__sa 3 3 3 3S%((//12 3 #3#78 3  3rcF<\SV`4\V4\PP ^4Vn\PPW^4Vn\PPW!^4Vn V!4Vn V!4Vn R#rm) rrr r!rhAdaptiveAvgPool2davgpoolrfc1fc2rr)r$rrrrr%s&&&&&rrSqueezeExcitation.__init__so D!xx11!4 88??>QG88??#3QG$, 0 2rc&<V^8dQhRS[RS[/#rinputr;r)rrs"rrrs,,F,v,rcVPV4pVPV4pVPV4pVPV4pVP V4#r)rrrrrr$rrEs&& r_scaleSqueezeExcitation._scalesI U#&$$U++rc&<V^8dQhRS[RS[/#rr)rrs"rrrsVrc4VPV4pW!,#r)rrs&& rrFSqueezeExcitation.forwards E"}r)rrrrr)rLrPrQrRrSr!rhr~SigmoidrrrFrTrUrVrWs@@rrrsG 6;XX]];@88;K;K 3 3,,rrcxaa]tRtRtoRtR]P PRRR3V3RlV3RllltRt Vt V;t #) MLPiaThis block implements the multi-layer perceptron (MLP) module. Args: in_channels (int): Number of channels of the input hidden_channels (List[int]): List of the hidden channel dimensions norm_layer (Callable[..., torch.nn.Module], optional): Norm layer that will be stacked on top of the linear layer. If ``None`` this layer won't be used. Default: ``None`` activation_layer (Callable[..., torch.nn.Module], optional): Activation function which will be stacked on top of the normalization layer (if not None), otherwise on top of the linear layer. If ``None`` this layer won't be used. Default: ``torch.nn.ReLU`` inplace (bool, optional): Parameter for the activation layer, which can optionally do the operation in-place. Default is ``None``, which uses the respective default values of the ``activation_layer`` and Dropout layer. bias (bool): Whether to use bias in the linear layer. Default ``True`` dropout (float): The probability for the dropout layer. Default: 0.0 NTgc<V^8dQhRS[RS[S[,RS[S[RS[P P 3,,RS[S[RS[P P 3,,RS[S[,RS[RS[/#) rr[hidden_channelsra.rbrdrdropout) rr2rrr!rhrir1r)rrs"rrMLP.__annotate__s"""c"Xc588??&:;< " #8C,@#AB " $ """rc d<Vf/MRV/p.p Tp VRRFp V P\PPWVR74VeV PV!V 44V PV!R/VB4V P\PP!V3/VB4T p K V P\PPWR,VR74V P\PP!V3/VB4\ S V`!V !\V4R#)Nrd)rr>rn)rvr!rhLinearDropoutrrr ) r$r[rrarbrdrrr{rzin_dim hidden_dimr%s &&&&&&&& rr MLP.__init__sY,@)#2.J MM%((//&4/H I% j45 MM*4V4 5 MM%((**7=f= >F /  ehhoofb.AoMN ehh&&w9&9: &!D!rrn) rLrPrQrRrSr!rhr~rrTrUrVrWs@@rrrs4 "@DEJXX]]"&"""rrcTaa]tRtRtoRtV3RlV3RlltV3RlRltRtVtV;t #)Permutei5zThis module returns a view of the tensor input with its dimensions permuted. Args: dims (List[int]): The desired ordering of dimensions c0<V^8dQhRS[S[,/#)rdims)r2r)rrs"rrPermute.__annotate__<sT#Yrc0<\SV`4WnR#r)rrr)r$rr%s&&rrPermute.__init__<s  rc&<V^8dQhRS[RS[/#r9r)rrs"rrr@s+++F+rcB\P!WP4#r)r!permuter)r$r:s&&rrFPermute.forward@s}}Q **r)r) rLrPrQrRrSrrFrTrUrVrWs@@rrr5s# +++rr)rwcollections.abcrtypingrrrr!rutilsr r rh functional interpolaterir SequentialrYrrrrrrnrrrs$,, 5hh!!-- 4T4Tn6,,6r/ -/ d. -. b$$N*"%((  *"Z +ehhoo +r