Ϧi5SSKJr SSKJr SSKJr SSKJr S9SjrS9Sjr\R\S 55r \R\\R"S 5S 555r \R\\R"S 5S:S j555r \R\S55r\S5r\S5r\S5r\S5r\S5r\S5r\S5r\R\\R,"SSSS9S;Sj555r\R\\R,"SSS9SS-j555r#\R\\RH"S.5S?S/j555r%\S05r&\R\\RH"S15S?S2j555r'\S@S3j5r(\SAS4j5r)\R\S\RT4SBS5jj55r+S6r,\R\SCS7j55r-\S85r.g)D) annotations)jit)core)mathczSnURnUS:aUS-nUS- nUS:aM[R"U5$)Nrrvaluer constexpr)ilog2ns W/var/www/html/ai-image-ml/venv/lib/python3.13/site-packages/triton/language/standard.py_log2r sC D A a% a   a% >>$ cnURn[R"XS- -S:H=(a US:g5$)Nrrr )r rs r_is_power_of_twors. A >>1A;1,7a 88rcX-S- U-$)z Computes the ceiling division of :code:`x` by :code:`div` :param x: the input number :type x: Block :param div: the divisor :type div: Block r)xdivs rcdivrs GaKC rsigmoidc<SS[R"U*5-- $)Nr)rexprs rrr+s DHHaRL !!rsoftmaxcU[US5- n[R"U5n[US5n[R"X4U5$)Nr)maxrrsumfdiv)r ieee_roundingznumdens rrr2s< C1I A ((1+C c1+C 99S} --rcB[R"XR/SS9$)z^ Returns a contiguous flattened view of :code:`x`. :param x: the input tensor :type x: Block T) can_reorder)rreshapenumelrs rravelr+<s <<GG9$ 77rc~X-U-nXC-nXV-nXt-n[R"X(- U5nXV-nXU--n XT-n X4$)a Transforms the indices of a row-major `size_i * size_j` matrix into the indices of a column-major matrix for each group of `size_g` rows. For example, for :code:`size_i = size_j = 4` and :code:`size_g = 2`, it will transform :: [[0 , 1 , 2 , 3 ], [4 , 5 , 6 , 7 ], [8 , 9 , 10, 11], [12, 13, 14, 15]] into :: [[0, 2, 4 , 6 ], [1, 3, 5 , 7 ], [8, 10, 12, 14], [9, 11, 13, 15]] rminimum) r jsize_isize_jsize_gijsize_gjgroup_idoff_inew_inew_js r swizzle2dr9Hs[, aBoG}H  E \\&.& 1F B K E LE <rc2[R"USU5$)a Returns a tensor filled with the scalar value 0 for the given :code:`shape` and :code:`dtype`. :param shape: Shape of the new array, e.g., (8, 16) or (8, ) :type shape: tuple of ints :param dtype: Data-type of the new array, e.g., :code:`tl.float16` :type dtype: DType r)rfull)shapedtypes rzerosr>ps 99UAu %%rcB[URUR5$)z{ Returns a tensor of zeros with the same shape and type as a given tensor. :param input: input tensor :type input: Tensor )r>r<r=)inputs r zeros_likerA}s ekk **rcU(aX:H=(a X:nOSnX:=(d Un[R"X`U5n[R"XaU5nXx4$NFrwhere) value1index1value2index2tie_break_lefttiegtv_reti_rets r_argmax_combinerOsM26?  CB JJr6 *E JJr6 *E <rc[XX#S5$NTrOrFrGrHrIs r_argmax_combine_tie_break_leftrT 664 @@rc[XX#S5$rCrRrSs r_argmax_combine_tie_break_fastrW 665 AArc.[R"X5$N)rmaximumabs r_elementwise_maxr_ << rr[return_indicesreturn_indices_tie_break_left)return_indices_arg tie_break_argNc[R"U5nU(a;U(a[R"X[US9$[R"X[US9$[R "UR R5[R "S5:a[R "UR R55(a UR[R5nOEUR R5(dS5eUR[R5n[R"X[US9$N keep_dims z"Expecting input to be integer type)r_promote_bfloat16_to_float32_reduce_with_indicesrTrWr r=primitive_bitwidth is_floatingtofloat32is_intint32reducer_r@axisrarbrhs rr r s  - -e 4E (,,U:Xdmn n,,U:Xdmn n >>%++88 9DNN2>%++88 9B >~~ekk55788.{{))++Q-QQ+,{{5(8INNrz minimum indexc [XSX#S9upEU$rv)rrws rargminrsTQ_ uFA Jrc X-$rZrr\s r _sum_combiner 5Lrr!cb[R"U5n[R"X[US9$)Nrg)rrjrrr)r@rtrhs rr!r!s'  - -e 4E ;;uLI FFrc X- $rZrr\s r _xor_combinerrrzxor sumc URRnUR5(d [S5e[R "XS9n[R "X[X#US9$)Nz#xor_sum only supported for integers)_builder)rhr _generator)typescalarrp ValueErrorrrjrrr)r@rtrhrr scalar_tys rxor_sumrsT !!I     >??  - -e GE ;;uLIeo pprcumsumcf[R"U5n[R"X[U5$rZ)rrjassociative_scanrr@rtreverses rrr%s)  - -e 4E  lG DDrc X-$rZrr\s r _prod_combiner1rrcumprodcf[R"U5n[R"X[U5$rZ)rrjrrrs rrr6s)  - -e 4E  mW EErcURU- nUSU--SSX2- S- -/n[R"X5n[R"SS5SSS2S4n[R"[ USU- -S5SS2SSS24U5R UR5n[R"[ Xg-S5SS2SSS24U5R UR5n [R"XR5n[R"XR5n [R"URRSS9n UR U SS9n U R U SS9n UR U SS9n U [R"X:U:gX- [U 55- nUR URSS9$)NrrrT)bitwidthsigned)bitcast) r*rr)arange broadcast_tor!rnr=r< get_int_dtyperlrErA)rflipr n_dimsn_outerr<ymaskleftrightidtypeileftirightixrys r_compare_and_swaprBsygg/G$q!t^QFJN0CDE QA ;;q! T1d] +D   Sa$h3AtQJ? G J J177 SD   c!(A.q$z:E B E Eagg NE <<gg &D LL (E   )C)CD QF GGFDG )E XXfdX +F fd #B tzz4># rc [R"[UR[ XR555 [R"[UR 55 [ UR 5n[ UR 5[ UR[ XR55- n[R"US/U-5n[R"XC5n[R"SS5SS2S4S[R"SS5- :Hn[R"X25H]nUn[R"SUS-5H*nX:wdM XS-:wdM[R"Xx5nM, [XG-US-SS9nM_ [R"X@R5nU$)z Flips a tensor `x` along the dimension `dim`. :param x: the first input tensor :type x: Block :param dim: the dimension to flip along (currently only final dimension supported) :type dim: int rrNrTrg) rrrr<rr*rr) expand_dimsrrr!) rrstepsstartrrr flip2r/s rrrsO ' c770K(LMN'01"!''NE!!''NU177=gg;V3W-XXE Qe $A "A KK1 ag &!dkk!Q.?*? ?D   u ,""1eai0Av!1u*((21  1q5D 1 - Q A Hrc[R"X5n[UR5S:XaU$[R"X"RSSSURS-/-5$)a Interleaves the values of two tensors along their last dimension. The two tensors must have the same shape. Equivalent to `tl.join(a, b).reshape(a.shape[:-1] + [2 * a.shape[-1]])` :param a: The first input tensor. :type a: Tensor :param b: The second input tensor. :type b: Tensor rNr)rjoinrr<r))r]r^cs r interleaversU !A 177|q ||Awws|q1772;.??@@r)r core.constexpr)F)NFTF)TFrC)NFNN)rF)r rrr)rrrrrr)rrrrrZ)/ __future__r runtime.jitrrrrr_tensor_member_fnr_add_math_1arg_docstrrrr+r9r>rArOrTrWr__add_reduction_docstrr rzrrrrrrrr!rbuiltinr_add_scan_docstrrrrrr CONSTEXPR_0rrrrrrrrsP"  9    I&"'"I&.'.88$$N & &++AABBI:J*IKOKO"O;KLM  AABBI:J*IKOKO"O;KLM E"G#G I&q'qx E!Ey!F"F))&  0"&TEUEU  0  8AAr