hFSSKrSSKrSSKJs Jr SSKJrJr SSKJr SS\S\ S\ S\ S \ S \ S \4S jjr SS\S\ S\ S\ S \ S \ S \4S jjr \RRS5 "SS\R 5r\RRS5 "SS\5rg)N)nnTensor)_log_api_usage_onceinputp block_sizeinplaceepstrainingreturnc|[RR5(d2[RR5(d[ [ 5 US:dUS:a[ SUS35eURS:wa[ SURS35eU(aUS:XaU$UR5upgp[X)U5nX-U -US-X- S -X- S ---- n [R"XgX- S -X- S -4URURS 9n U RU 5 [R"XS-/S-S S 9n [R "U S X"4US-S9n S U - n U R#5XKR%5-- n U(a"UR'U 5R'U 5 U$X -U -nU$)a Implements DropBlock2d from `"DropBlock: A regularization method for convolutional networks" `. Args: input (Tensor[N, C, H, W]): The input tensor or 4-dimensions with the first one being its batch i.e. a batch with ``N`` rows. p (float): Probability of an element to be dropped. block_size (int): Size of the block to drop. inplace (bool): If set to ``True``, will do this operation in-place. Default: ``False``. eps (float): A value added to the denominator for numerical stability. Default: 1e-6. training (bool): apply dropblock if is ``True``. Default: ``True``. Returns: Tensor[N, C, H, W]: The randomly zeroed tensor after dropblock. ?4drop probability has to be between 0 and 1, but got .z#input should be 4 dimensional. Got dimensions.rdtypedevicervalue)rrstride kernel_sizepadding)torchjit is_scripting is_tracingr drop_block2d ValueErrorndimsizeminemptyrr bernoulli_Fpad max_pool2dnumelsummul_) rrr r r r NCHWgammanoisenormalize_scales R/var/www/auris/envauris/lib/python3.13/site-packages/torchvision/ops/drop_block.pyr#r# s& 99 ! ! # #EII,@,@,B,BL)3w!c'OPQsRSTUU zzQ>uzzl,WXX qCx JA!ZA&J UQYJMq~/AanWXFX.YZ [E KKq~11>A3EFekkbgbnbn oE U EE%/*Q.a 8E LLvJ;S]gkl]l mE IEkkmsYY['89O 5/ L / Lc[RR5(d2[RR5(d[ [ 5 US:dUS:a[ SUS35eURS:wa[ SURS35eU(aUS:XaU$UR5upgpn [X(X5nX-U -U -US-X- S -X- S --X- S ---- n [R"XgX- S -X- S -X- S -4URURS 9n U RU 5 [R"XS -/S -S S9n [R "U SX"U4US -S9n S U - n U R#5XLR%5-- n U(a"UR'U 5R'U 5 U$X -U -nU$)a Implements DropBlock3d from `"DropBlock: A regularization method for convolutional networks" `. Args: input (Tensor[N, C, D, H, W]): The input tensor or 5-dimensions with the first one being its batch i.e. a batch with ``N`` rows. p (float): Probability of an element to be dropped. block_size (int): Size of the block to drop. inplace (bool): If set to ``True``, will do this operation in-place. Default: ``False``. eps (float): A value added to the denominator for numerical stability. Default: 1e-6. training (bool): apply dropblock if is ``True``. Default: ``True``. Returns: Tensor[N, C, D, H, W]: The randomly zeroed tensor after dropblock. rrrrz#input should be 5 dimensional. Got rrrrrr)rrrr)rr r!r"r drop_block3dr$r%r&r'r(rrr)r*r+ max_pool3dr-r.r/)rrr r r r r0r1Dr2r3r4r5r6s r7r=r=7s& 99 ! ! # #EII,@,@,B,BL)3w!c'OPQsRSTUU zzQ>uzzl,WXX qCx JJLMA!ZA)J UQY] A 1>A3E!.[\J\2]abaorsas2tu vE KK q~!1>A#5q~7IJRWR]R]fkfrfr E U EE%/*Q.a 8E LL ijj-Q[eij[j E IEkkmsYY['89O 5/ L / Lr8r#c h^\rSrSrSrSS\S\S\S\SS4 U4S jjjrS \ S\ 4S jr S\ 4S jr S r U=r$) DropBlock2dlz See :func:`drop_block2d`. rr r r r NcR>[TU]5 XlX lX0lX@lgN)super__init__rr r r selfrr r r __class__s r7rFDropBlock2d.__init__qs" $ r8rc[XRURURURUR 5$z Args: input (Tensor): Input feature map on which some areas will be randomly dropped. Returns: Tensor: The tensor after DropBlock layer. )r#rr r r r rHrs r7forwardDropBlock2d.forwardy.E664??DLL$((TXTaTabbr8cURRSURSURSURS3nU$)Nz(p=z , block_size=z , inplace=))rI__name__rr r )rHss r7__repr__DropBlock2d.__repr__sC~~&&'s466(-?PPZ[_[g[gZhhi jr8)r r r rFư>)rS __module__ __qualname____firstlineno____doc__floatintboolrFrrNstrrU__static_attributes__ __classcell__rIs@r7rArAls]%S4e`dcVcc#r8rAr=c Z^\rSrSrSrS S\S\S\S\SS4 U4S jjjrS \ S\ 4S jr S r U=r $) DropBlock3dz See :func:`drop_block3d`. rr r r r Nc&>[TU]XX45 grD)rErFrGs r7rFDropBlock3d.__init__s 5r8rc[XRURURURUR 5$rL)r=rr r r r rMs r7rNDropBlock3d.forwardrPr8rW) rSrYrZr[r\r]r^r_rFrrNrarbrcs@r7reresR6%6S646e6`d66cVcccr8re)FrXT)rtorch.fxtorch.nn.functionalr functionalr*rutilsrr]r^r_r#r=fxwrapModulerArerkr8r7rss 'ko+ ++),+7;+JO+cg+ +^ko/ //),/7;/JO/cg/ /d n"))8 nc+cr8