+ NV-j84^RIt^RIHt^RIHtHtHt^RIHt ^RI H t Rt Rt Rt!RR ] 4t!R R ]4t!R R ]4t!RR]4t!RR]4t!RR]4t!RR]4t!RR]4t!RR]4t!RR]4tR#)N) accumulate)OptionalTupleUnion)Modulec\V\\34'd'\V4V8wd \ V4h\V4#\V\ 4'g \ V4hV.V,#)N) isinstancelisttuplelen ValueErrorint)xnmsgs&&&f/Volumes/fast/ai/experiments/ui-tars-smoke/.venv/lib/python3.14/site-packages/mlx/nn/layers/pooling.py_value_or_listr sS!dE]## q6Q;S/ !Aw a  o 37NcxV^,.p\VR,V4F-wrEVPWE,4VPV4K/ VPVR,4\V4^, p^.\^V^4O\^V^4OVNpVP V4pVP V4pV#)rNN)zipappendr rangereshape transpose)rshape window_shape new_shapesw last_axis axis_orders&&& r _non_overlapping_sliding_windowsr%sq IE"I|, -U2YI"IQeAy!,QuQ 1/EQyQJ )A JA Hrc 6VP^8d\RVP R24hVP^R p\V4\V4u;8Xd\V48Xg3M\R\V4 R\V4 R\V4 R24hVPp\;QJd&R\ W1V44F 'dK RM R M!R\ W1V444'd \ WV4#\\\\\VR,4\P4444R ,pV^,.pT\ W1V4UUU u.uFwrxp Wx, V ,^,NK up pp, pWa, pWdR ,., pVR ,p T \ V^R V4U U u.uF wrW,NK up p , p W^R , p WR R , p \P!WV 4#uup ppiuup p i)zcTo extract sliding windows at least 1 spatial dimension (3 total) is needed but the input only has z dimensions.zTo extract sliding windows the window shapes and strides must have the same number of spatial dimensions as the signal but the signal has z dims and the window shape has z and strides have .c3V"TFwrpW#8H;'d W,^8HxK! R#5irN).0sizewindowstrides& r #_sliding_windows..8s1 $S D& //T]a//$Ss))FTr:NrNNr)r)ndimr rr allrr%r reversedroperatormulmx as_strided) rrwindow_strides spatial_dimsrstrides final_shaper-r.r/ final_strides og_strides &&& r_sliding_windowsr?'svvz::;&& O  771R=L  \!2 Ic.6I I|$%%DSEVDWX #N 34A 7   GGE s $' N$S sss $' N$S 0,GG8DHUT\,BHLL!QRSTUWXG8*K$' N$S$S D& 6!A%%$SKK"I;KBKM47" ~4V4V0y 4VMQr]"MRS\!M == 77s #"H Hc>aa]tRt^TtoV3RltRtRtRtVtV;t #)_Poolc<\SV`4WnW nW0nW@nWPn\\\VP4)^, R^44Vn R#)rNr) super__init___pooling_function _kernel_size_stride_padding_padding_valuer rr _axes)selfpooling_function kernel_sizer/padding padding_value __class__s&&&&&&rrD_Pool.__init__UsR !1'  +5#d&7&7"8!81!.bs/AQ44sz kernel_size=z , stride=z , padding=)r rFrGrH)rKksstpds& r _extra_repr_Pool._extra_repr_s_ 4$$ % 4<<  U//UU// /bT2$j==rc\;QJd&RVP4F 'gK RM RM!RVP44'd>\P!VR.VP,R.,VPR7p\ WP VP4pVPWP4#)c38"TFq^,^8xK R#5ir*r+rTs& rr0!_Pool.__call__..gs/AtaxsTF)constant_values)rr) anyrHr7padrIr?rFrGrErJ)rKrs&&r__call___Pool.__call__fs 3//333// / /4==(F83 $ 3 3A Q 1 14<< @%%a44r)rJrFrHrIrErG) __name__ __module__ __qualname____firstlineno__rDrZrb__static_attributes____classdictcell__ __classcell__rP __classdict__s@@rrArATsF>55rrAcBaa]tRt^qtoRV3RlV3RllltRtVtV;t#)_Pool1dc <V^8dQhRS[S[S[S[,3,RS[S[S[S[S[,3,,RS[S[S[S[,3,/#rMr/rNrrrr)formatrls"r __annotate___Pool1d.__annotate__rs\XX3c ?+ X sE#J/0 X sE#J' XrcB<\V4PpRp\V^VPVR44pVe\V^VPVR44pMTp\V^VPVR44pVUu.uFqV3NK pp\S V`WWEV4R#uupi)z<[{}] '{}' must be an integer or a tuple containing 1 integerrMNr/rNtyperdrrsrCrD rKrLrOrMr/rN class_namerrUrPs &&&&&& rrD_Pool1d.__init__rs$Z(( L$ CJJz=A   #FAszz*h/OPF F !SZZ I-NO#*+7aq67+ )W,; Br+NrrdrerfrgrDrhrirjrks@@rrnrnqXXXrrncBaa]tRt^toRV3RlV3RllltRtVtV;t#)_Pool2dc <V^8dQhRS[S[S[S[S[3,3,RS[S[S[S[S[S[3,3,,RS[S[S[S[S[S[3,3,,/#rprr)rsrls"rrt_Pool2d.__annotate__spXX3c3h/0 X sE#s(O345 X %U38_ 456 XrcB<\V4PpRp\V^VPVR44pVe\V^VPVR44pMTp\V^VPVR44pVUu.uFqV3NK pp\S V`WWEV4R#uupi)z=[{}] '{}' must be an integer or a tuple containing 2 integersrMNr/rNrwrys &&&&&& rrD_Pool2d.__init__$Z(( M$ CJJz=A   #FAszz*h/OPF F !SZZ I-NO#*+7aq67+ )W,r|r+r}r~rks@@rrrrrrcBaa]tRt^toRV3RlV3RllltRtVtV;t#)_Pool3dc <V^8dQhRS[S[S[S[S[S[3,3,RS[S[S[S[S[S[S[3,3,,RS[S[S[S[S[S[S[3,3,,/#rprr)rsrls"rrt_Pool3d.__annotate__syXX3c3m 445 X sE#sC-$889: X %U3S=%9 9:; XrcB<\V4PpRp\V^VPVR44pVe\V^VPVR44pMTp\V^VPVR44pVUu.uFqV3NK pp\S V`WWEV4R#uupi)z=[{}] '{}' must be an integer or a tuple containing 3 integersrMNr/rNrwrys &&&&&& rrD_Pool3d.__init__rr|r+r}r~rks@@rrrrrrcFaa]tRt^toRtRV3RlV3RllltRtVtV;t#) MaxPool1daApplies 1-dimensional max pooling. Spatially downsamples the input by taking the maximum of a sliding window of size ``kernel_size`` and sliding stride ``stride``. Args: kernel_size (int or tuple(int)): The size of the pooling window kernel. stride (int or tuple(int), optional): The stride of the pooling window. Default: ``kernel_size``. padding (int or tuple(int), optional): How much negative infinity padding to apply to the input. The padding amount is applied to both sides of the spatial axis. Default: ``0``. Examples: >>> import mlx.core as mx >>> import mlx.nn.layers as nn >>> x = mx.random.normal(shape=(4, 16, 5)) >>> pool = nn.MaxPool1d(kernel_size=2, stride=2) >>> pool(x) c <V^8dQhRS[S[S[S[,3,RS[S[S[S[S[,3,,RS[S[S[S[,3,/#rprr)rsrls"rrtMaxPool1d.__annotate__sZNN3c ?+NsE#J/0NsE#J' Nrc\<\SV`\P\ R4)WV4R#infNrCrDr7maxfloatrKrMr/rNrPs&&&&rrDMaxPool1d.__init__" %, WMrr+r} rdrerfrg__doc__rDrhrirjrks@@rrrs*NNNrrcFaa]tRt^toRtRV3RlV3RllltRtVtV;t#) AvgPool1daApplies 1-dimensional average pooling. Spatially downsamples the input by taking the average of a sliding window of size ``kernel_size`` and sliding stride ``stride``. Args: kernel_size (int or tuple(int)): The size of the pooling window kernel. stride (int or tuple(int), optional): The stride of the pooling window. Default: ``kernel_size``. padding (int or tuple(int), optional): How much zero padding to apply to the input. The padding amount is applied to both sides of the spatial axis. Default: ``0``. Examples: >>> import mlx.core as mx >>> import mlx.nn.layers as nn >>> x = mx.random.normal(shape=(4, 16, 5)) >>> pool = nn.AvgPool1d(kernel_size=2, stride=2) >>> pool(x) c <V^8dQhRS[S[S[S[,3,RS[S[S[S[S[,3,,RS[S[S[S[,3,/#rprr)rsrls"rrtAvgPool1d.__annotate__sZCC3c ?+CsE#J/0CsE#J' CrcH<\SV`\P^WV4R#r*rCrDr7meanrs&&&&rrDAvgPool1d.__init__ !['Brr+r}rrks@@rrrs*CCCrrcFaa]tRt^toRtRV3RlV3RllltRtVtV;t#) MaxPool2da1Applies 2-dimensional max pooling. Spatially downsamples the input by taking the maximum of a sliding window of size ``kernel_size`` and sliding stride ``stride``. The parameters ``kernel_size``, ``stride``, and ``padding`` can either be: * a single ``int`` -- in which case the same value is used for both the height and width axis. * a ``tuple`` of two ``int`` s -- in which case, the first ``int`` is used for the height axis, the second ``int`` for the width axis. Args: kernel_size (int or tuple(int, int)): The size of the pooling window. stride (int or tuple(int, int), optional): The stride of the pooling window. Default: ``kernel_size``. padding (int or tuple(int, int), optional): How much negative infinity padding to apply to the input. The padding is applied on both sides of the height and width axis. Default: ``0``. Examples: >>> import mlx.core as mx >>> import mlx.nn.layers as nn >>> x = mx.random.normal(shape=(8, 32, 32, 4)) >>> pool = nn.MaxPool2d(kernel_size=2, stride=2) >>> pool(x) c <V^8dQhRS[S[S[S[S[3,3,RS[S[S[S[S[S[3,3,,RS[S[S[S[S[S[3,3,,/#rprr)rsrls"rrtMaxPool2d.__annotate__snNN3c3h/0NsE#s(O345N%U38_ 456 Nrc\<\SV`\P\ R4)WV4R#rrrs&&&&rrDMaxPool2d.__init__rrr+r}rrks@@rrrs8NNNrrcFaa]tRtRtoRtRV3RlV3RllltRtVtV;t#) AvgPool2dia(Applies 2-dimensional average pooling. Spatially downsamples the input by taking the average of a sliding window of size ``kernel_size`` and sliding stride ``stride``. The parameters ``kernel_size``, ``stride``, and ``padding`` can either be: * a single ``int`` -- in which case the same value is used for both the height and width axis. * a ``tuple`` of two ``int`` s -- in which case, the first ``int`` is used for the height axis, the second ``int`` for the width axis. Args: kernel_size (int or tuple(int, int)): The size of the pooling window. stride (int or tuple(int, int), optional): The stride of the pooling window. Default: ``kernel_size``. padding (int or tuple(int, int), optional): How much zero padding to apply to the input. The padding is applied on both sides of the height and width axis. Default: ``0``. Examples: >>> import mlx.core as mx >>> import mlx.nn.layers as nn >>> x = mx.random.normal(shape=(8, 32, 32, 4)) >>> pool = nn.AvgPool2d(kernel_size=2, stride=2) >>> pool(x) c <V^8dQhRS[S[S[S[S[3,3,RS[S[S[S[S[S[3,3,,RS[S[S[S[S[S[3,3,,/#rprr)rsrls"rrtAvgPool2d.__annotate__:snCC3c3h/0CsE#s(O345C%U38_ 456 CrcH<\SV`\P^WV4R#r*rrs&&&&rrDAvgPool2d.__init__:rrr+r}rrks@@rrrs8CCCrrcFaa]tRtRtoRtRV3RlV3RllltRtVtV;t#) MaxPool3diCa|Applies 3-dimensional max pooling. Spatially downsamples the input by taking the maximum of a sliding window of size ``kernel_size`` and sliding stride ``stride``. The parameters ``kernel_size``, ``stride``, and ``padding`` can either be: * a single ``int`` -- in which case the same value is used for the depth, height, and width axis. * a ``tuple`` of three ``int`` s -- in which case, the first ``int`` is used for the depth axis, the second ``int`` for the height axis, and the third ``int`` for the width axis. Args: kernel_size (int or tuple(int, int, int)): The size of the pooling window. stride (int or tuple(int, int, int), optional): The stride of the pooling window. Default: ``kernel_size``. padding (int or tuple(int, int, int), optional): How much negative infinity padding to apply to the input. The padding is applied on both sides of the depth, height and width axis. Default: ``0``. Examples: >>> import mlx.core as mx >>> import mlx.nn.layers as nn >>> x = mx.random.normal(shape=(8, 16, 32, 32, 4)) >>> pool = nn.MaxPool3d(kernel_size=2, stride=2) >>> pool(x) c <V^8dQhRS[S[S[S[S[S[3,3,RS[S[S[S[S[S[S[3,3,,RS[S[S[S[S[S[S[3,3,,/#rprr)rsrls"rrtMaxPool3d.__annotate__aswNN3c3m 445NsE#sC-$889:N%U3S=%9 9:; Nrc\<\SV`\P\ R4)WV4R#rrrs&&&&rrDMaxPool3d.__init__arrr+r}rrks@@rrrCs:NNNrrcFaa]tRtRtoRtRV3RlV3RllltRtVtV;t#) AvgPool3dijasApplies 3-dimensional average pooling. Spatially downsamples the input by taking the average of a sliding window of size ``kernel_size`` and sliding stride ``stride``. The parameters ``kernel_size``, ``stride``, and ``padding`` can either be: * a single ``int`` -- in which case the same value is used for the depth, height, and width axis. * a ``tuple`` of three ``int`` s -- in which case, the first ``int`` is used for the depth axis, the second ``int`` for the height axis, and the third ``int`` for the width axis. Args: kernel_size (int or tuple(int, int, int)): The size of the pooling window. stride (int or tuple(int, int, int), optional): The stride of the pooling window. Default: ``kernel_size``. padding (int or tuple(int, int, int), optional): How much zero padding to apply to the input. The padding is applied on both sides of the depth, height and width axis. Default: ``0``. Examples: >>> import mlx.core as mx >>> import mlx.nn.layers as nn >>> x = mx.random.normal(shape=(8, 16, 32, 32, 4)) >>> pool = nn.AvgPool3d(kernel_size=2, stride=2) >>> pool(x) c <V^8dQhRS[S[S[S[S[S[3,3,RS[S[S[S[S[S[S[3,3,,RS[S[S[S[S[S[S[3,3,,/#rprr)rsrls"rrtAvgPool3d.__annotate__swCC3c3m 445CsE#sC-$889:C%U3S=%9 9:; CrcH<\SV`\P^WV4R#r*rrs&&&&rrDAvgPool3d.__init__rrr+r}rrks@@rrrjs:CCCrr)r5 itertoolsrtypingrrrmlx.corecorer7mlx.nn.layers.baserrr%r?rArnrrrrrrrrr+rrrs ))%   *8Z5F5:XeX0XeX0XeX0NN>CC>#N#NL#C#CL$N$NN$C$Cr