+ j/&^RIt^RIHt^RIt^RIt^RIHtHt^RIHt^RI H t ^RI H t ^RI HtHt^RIHt^R IHtHtR tR tR t]!R R7R4t]P2P4RRRll4t!RR]P84tR#)N)Union)nnTensor)is_compile_supported)BroadcastingList2)_pair)_assert_has_ops_has_ops)_log_api_usage_once)check_roi_boxes_shapeconvert_boxes_to_roi_formatc aV3RlpV#)zcLazily wrap a function with torch.compile on the first call This avoids eagerly importing dynamo. cJ<a\P!S4VV3Rl4pV#)c<\P!S3/SBp\P!S4!V4\ 4SP &V!V/VB#N)torchcompile functoolswrapsglobals__name__)argskwargs compiled_fncompile_kwargsfns*, Z/home/luca/screen-checking/.venv/lib/python3.14/site-packages/torchvision/ops/roi_align.py compile_hook7lazy_compile..decorate_fn..compile_hooksD--=n=K%.__R%8%EGIbkk "// /)rr)rrrsf r decorate_fn!lazy_compile..decorate_fns&   0  0 r )rr!sl r lazy_compiler$s  r cLaaaaaSP4wporxVP^R7pVP^R7pVP4p VP4p \P!W^, 8V^, V ^,4p \P!W^, 8V^, V 4p \P!W^, 8VP SP 4V4p\P!W^, 8V^, V ^,4p \P!W^, 8V^, V 4p \P!W^, 8VP SP 4V4pW), p W:, pRV , pRV, pVVVVV3RlpV!W4pV!W4pV!W4pV!W4pRpV!VV4pV!W4pV!V V4pV!W4pVV,VV,,VV,,VV,,pV#)rmin?c&<SeESfQh\P!SR,V^4p\P!SR,V^4pSSR,\P!SSPR7R,VR,VR,3,#)Ndevice)NNNNr,)r,NNNNN)Nr,NNNNr,Nr,Nr,Nr,NNr,Nr,)rwherearanger+)yxchannelsinput roi_batch_indxmaskymasks&&r masked_index+_bilinear_interpolate..masked_indexBs  $ $$ E*-q!4A E*-q!4A 9 : LL%,, 78W X ' ( ' ( *  r c0VR,VR,,#)r,r-r.r#)r1r2s&&r outer_prod)_bilinear_interpolate..outer_prodWs*+a0I.JJJr )sizeclampintrr/todtype)r4r5r1r2r7r6_heightwidthy_lowx_lowy_highx_highlylxhyhxr8v1v2v3v4r;w1w2w3w4valr3sff&&ff @r_bilinear_interpolaterV#s"'Ax AA AA EEGE EEGE [[1*,fqj%!) DF KK!+VaZ ?E EaZ'ekk):A>A [[!)+UQY BF KK*EAIu =E EQY&U[[(91=A B B rB rB    e #B e $B f $B f %BK B B B B B B B B r'BG b2g %R /C Jr c\P!4'dBVP'd0VP\P8wdVP 4#V#r)ris_autocast_enabledis_cudarAdoublefloat)tensors&r maybe_castr]es9   ""v~~~&,,%,,:V||~ r T)dynamicc&VPp\V4p\V4pVP4wrp \P!W0P R7p \P!W@P R7p VR,P 4p V'dRMRpVR,V,V, pVR ,V,V, pVR ,V,V, pVR ,V,V, pVV, pVV, pV'g1\P!VRR7p\P!VRR7pVV, pVV, pV^8pV'dTM\P!VV, 4pV'dTM\P!VV, 4pV'd]\VV,^4p\P!VVP R7p\P!VVP R7pRpRpM\P!VV,^R7p\P!WP R7p\P!WP R7pVR ,VR ,8pVR ,VR ,8pRpV!V4V R,V!V4,,VR,R,PVP4V!VV, 4,,p V!V4V R,V!V4,,VR,R,PVP4V!VV, 4,,p!\W V V!VV4p"V'g?\P!VR,V"^4p"\P!VR,V"^4p"V"PR4p#\V\P4'dV#VR,,p#M V#V,p#V#PV4p#V##)r*g?gr(r&NcVR,#)r,)r,NNr#)ts&rfrom_K_roi_align..from_Ks r )r,r)r,)r,)r,)r,)Nr,)r,N)Nr,N)NNr,)r,NNNr,N)r,NNNNr,))r,NNN)rAr]r=rr0r+r?r>ceilmaxr@rVr/sum isinstancer)$r4rois spatial_scale pooled_height pooled_widthsampling_ratioaligned orig_dtyperBrCrDphpwr5offset roi_start_w roi_start_h roi_end_w roi_end_h roi_width roi_height bin_size_h bin_size_wexact_samplingroi_bin_grid_hroi_bin_grid_wcountiyixr7r6rbr1r2rUoutputs$&&&&&&& r _roi_alignrrs6J u E d D**,Aq% mLL 9B l<< 8B JNN$MSFt*},v5Kt*},v5KT ]*V3IT ]*V3IK'I[(J KK s3 [[5 m+J\)J#a'N'5^5::jS`F`;aN'5^5::iR^F^;_NN^3Q7 \\. > \\. > N^;C \\& 6 \\% 57 nW557 nW55  { ] fZ0 0 1 m s " & &u{{ 3fZ.=X6Y Y Z { ] fZ0 0 1 m s " & &u{{ 3fZ.=X6Y Y Z aE5 IC kk% <=sAFkk% <=sAF WWX F%&&%+,,% YYz "F Mr cV^8dQhR\R\\\\,3,R\\,R\ R\R\ R\/#)rer4boxes output_sizerorrrsreturn)rrlistrr?r[bool)formats"r __annotate__rsjQQ Q f% &Q#3'Q Q  Q  Q Qr c v\PP4'g4\PP4'g\ \ 4\ V4Tp\V4p\V\P4'g \V4p\PP4'gVP'd\4VP4pVP4pVP4p VP'g \!R4hVP4p VP4p VP4p \P"P$P'V V VVV V VV^,V^,VV4 p \P(!WVR7#\PP4'g\+4'dR\P,!4'dxVP.'g%VP0'gVP2'dB\5VP6P84'd\;WW2^,V^,WE4#\4\P"P$P WW2^,V^,WE4#)a Performs Region of Interest (RoI) Align operator with average pooling, as described in Mask R-CNN. Args: input (Tensor[N, C, H, W]): The input tensor, i.e. a batch with ``N`` elements. Each element contains ``C`` feature maps of dimensions ``H x W``. If the tensor is quantized, we expect a batch size of ``N == 1``. boxes (Tensor[K, 5] or List[Tensor[L, 4]]): the box coordinates in (x1, y1, x2, y2) format where the regions will be taken from. The coordinate must satisfy ``0 <= x1 < x2`` and ``0 <= y1 < y2``. If a single Tensor is passed, then the first column should contain the index of the corresponding element in the batch, i.e. a number in ``[0, N - 1]``. If a list of Tensors is passed, then each Tensor will correspond to the boxes for an element i in the batch. output_size (int or Tuple[int, int]): the size of the output (in bins or pixels) after the pooling is performed, as (height, width). spatial_scale (float): a scaling factor that maps the box coordinates to the input coordinates. For example, if your boxes are defined on the scale of a 224x224 image and your input is a 112x112 feature map (resulting from a 0.5x scaling of the original image), you'll want to set this to 0.5. Default: 1.0 sampling_ratio (int): number of sampling points in the interpolation grid used to compute the output value of each pooled output bin. If > 0, then exactly ``sampling_ratio x sampling_ratio`` sampling points per bin are used. If <= 0, then an adaptive number of grid points are used (computed as ``ceil(roi_width / output_width)``, and likewise for height). Default: -1 aligned (bool): If False, use the legacy implementation. If True, pixel shift the box coordinates it by -0.5 for a better alignment with the two neighboring pixel indices. This version is used in Detectron2 Returns: Tensor[K, C, output_size[0], output_size[1]]: The pooled RoIs. z3If input is quantized, rois must also be quantized.)scale zero_point)rjit is_scripting is_tracingr roi_alignr rrmrr is_quantizedr q_scale q_zero_pointint_repr ValueErrorops torchvision qroi_align!_make_per_tensor_quantized_tensorr $are_deterministic_algorithms_enabledrYis_mpsis_xpurr+typer)r4rrrorrrsrn input_scaleinput_zero_pointint_data rois_scalerois_zero_pointrois_intout_ints&&&&&& rrrsR 99 ! ! # #EII,@,@,B,BI&% D $K dELL ) )*40 99 ! ! # #(:(:(:mmo  --/>>#   RS S\\^ ++-==?))''22        N N    66w^noo 99 ! ! # # ::<<%---SXS_S_S_chcococo"5<<#4#455e=a.+VW.Zhr r 99 * * ]NKNN r cjaa]tRtRtoRtR V3RlV3RllltV3RlRltV3RlRltR tVt V;t #) RoIAligni z See :func:`roi_align`. cB<V^8dQhRS[S[,RS[RS[RS[/#)rerrorrrs)rr?r[r)r __classdict__s"rrRoIAlign.__annotate__%s8  &s+     r cj<\SV`4\V4WnW nW0nW@nR#r)super__init__r rrorrrs)selfrrorrrs __class__s&&&&&rrRoIAlign.__init__%s/ D!&*, r cR<V^8dQhRS[RS[S[S[S[,3,RS[/#)rer4rnr)rrr)rrs"rrr3s5ooVo5f1E+Fo6or cp\WVPVPVPVP4#r)rrrorrrs)rr4rns&&&rforwardRoIAlign.forward3s.d&6&68J8JDL_L_aeamamnnr c <V^8dQhRS[/#)rer)str)rrs"rrr6s  # r c VPP RVP RVP RVP RVP R2 pV#)z (output_size=z, spatial_scale=z, sampling_ratio=z , aligned=))rrrrorrrs)rss& r__repr__RoIAlign.__repr__6s^~~&&'(++,t112 3 34   r )rsrrrro)F) r __module__ __qualname____firstlineno____doc__rrr__static_attributes____classdictcell__ __classcell__)rrs@@rrr s/  oo   r r)r(rhF)rtypingrrtorch.fxrrtorch._dynamo.utilsrtorch.jit.annotationsrtorch.nn.modules.utilsrtorchvision.extensionr r utilsr _utilsr r r$rVr]rfxwraprModulerr#r rrs 43(;'F&=DdUUpQQhryyr