ΑipJf%SSKJr SSKJrJr SSKrSSKJr SSK J r SSK J r J r Jr SSKJr SSKJr SS KJr S S KJrJrJr \(aSS K Jr \S rS\S'\SrS\S'/rSSSjjrSSSjjrSSSjjr g)) annotations) TYPE_CHECKINGLiteralN) TypeAlias)_C_ops) check_dtype check_typecheck_variable_and_dtype)Variable) LayerHelper)in_dynamic_or_pir_mode)convert_out_size_to_listget_out_size_tensor_inputsreshape_lhs_rhs)Tensorsummeanmaxminr _ReduceOpaddsubmuldiv _MessageOpcUS;a[SU35e[5(a2[US5n[R"XX#R 5U5$[ USSS5 [ USSS5 [ USSS5 U(a6[US [[R[R[4S5 [U[5(a[URS S S /S5 [!S0[#5D6nUR%URS 9nUR%S S S9nXUS.n SUR 50n ['U U USS9 UR)SU XxS.U S9 U$)aG Graph Learning message passing api. This api is mainly used in Graph Learning domain, and the main purpose is to reduce intermediate memory consumption in the process of message passing. Take `x` as the input tensor, we first use `src_index` to gather the corresponding data, and then use `dst_index` to update the corresponding position of output tensor in different reduce ops, like sum, mean, max, or min. Besides, we can use `out_size` to set necessary output shape. .. code-block:: text Given: x = [[0, 2, 3], [1, 4, 5], [2, 6, 7]] src_index = [0, 1, 2, 0] dst_index = [1, 2, 1, 0] reduce_op = "sum" out_size = None Then: out = [[0, 2, 3], [2, 8, 10], [1, 4, 5]] Args: x (Tensor): The input tensor, and the available data type is float32, float64, int32, int64. And we support float16 in gpu version. src_index (Tensor): An 1-D tensor, and the available data type is int32, int64. dst_index (Tensor): An 1-D tensor, and should have the same shape as `src_index`. The available data type is int32, int64. reduce_op (str): Different reduce ops, including `sum`, `mean`, `max`, `min`. Default value is `sum`. out_size (int|Tensor|None): We can set `out_size` to get necessary output shape. If not set or out_size is smaller or equal to 0, then this input will not be used. Otherwise, `out_size` should be equal with or larger than max(dst_index) + 1. name (str, optional): Name for the operation (optional, default is None). For more information, please refer to :ref:`api_guide_Name`. Returns: - out (Tensor), the output tensor, should have the same shape and same dtype as input tensor `x`. If `out_size` is set correctly, then it should have the same shape as `x` except the 0th dimension. Examples: .. code-block:: python >>> import paddle >>> x = paddle.to_tensor([[0, 2, 3], [1, 4, 5], [2, 6, 7]], dtype="float32") >>> indexes = paddle.to_tensor([[0, 1], [1, 2], [2, 1], [0, 0]], dtype="int32") >>> src_index, dst_index = indexes[:, 0], indexes[:, 1] >>> out = paddle.geometric.send_u_recv(x, src_index, dst_index, reduce_op="sum") >>> print(out.numpy()) [[ 0. 2. 3.] [ 2. 8. 10.] [ 1. 4. 5.]] >>> x = paddle.to_tensor([[0, 2, 3], [1, 4, 5], [2, 6, 7]], dtype="float32") >>> indexes = paddle.to_tensor([[0, 1], [2, 1], [0, 0]], dtype="int32") >>> src_index, dst_index = indexes[:, 0], indexes[:, 1] >>> out_size = paddle.max(dst_index) + 1 >>> out = paddle.geometric.send_u_recv(x, src_index, dst_index, reduce_op="sum", out_size=out_size) >>> print(out.numpy()) [[ 0. 2. 3.] [ 2. 8. 10.]] >>> x = paddle.to_tensor([[0, 2, 3], [1, 4, 5], [2, 6, 7]], dtype="float32") >>> indexes = paddle.to_tensor([[0, 1], [2, 1], [0, 0]], dtype="int32") >>> src_index, dst_index = indexes[:, 0], indexes[:, 1] >>> out = paddle.geometric.send_u_recv(x, src_index, dst_index, reduce_op="sum") >>> print(out.numpy()) [[ 0. 2. 3.] [ 2. 8. 10.] [ 0. 0. 0.]] r@reduce_op should be `sum`, `mean`, `max` or `min`, but received graph_send_recvXfloat32float64int32int64float16 Src_indexr&r' Dst_indexout_sizer&r'dtypeTr. stop_gradient)r"r)r+ reduce_opinputsattrsr,op_typeOut Dst_counttyper3outputsr4) send_u_recv) ValueErrorr rrr<upperr r intnpr&r'r isinstancerr.r locals"create_variable_for_type_inferencer append_op) x src_index dst_indexr1r,namehelperout dst_countr3r4s j/var/www/html/banglarbhumi/venv/lib/python3.13/site-packages/paddle/geometric/message_passing/send_recv.pyr<r<7sv55Nyk Z  +H6GH!! )__%6   !  ?   ! {$68I  ! {$68I   bhh(3!   h ) ) '"!  7fh777agg7F==> yIioo/0"%   "8   c US;a[SU35eUS;a[SU35e[X5upUS:XaSnU*nUS:Xa SnS US -- n[5(aD[US 5n[R "UUUUUR 5UR 5U5$[US S S 5 [USS S 5 [USSS 5 [USSS 5 U(a6[US[[R[R[4S 5 [U[5(a[UR SSS/S 5 [#S0[%5D6nUR'UR S9n UR'SSS9n UUUUS.n UR 5UR 5S.n [)U U US S9 UR+S U XS.U S9 U $)a Graph Learning message passing api. This api is mainly used in Graph Learning domain, and the main purpose is to reduce intermediate memory consumption in the process of message passing. Take `x` as the input tensor, we first use `src_index` to gather the corresponding data, after computing with `y` in different message ops like add/sub/mul/div, then use `dst_index` to update the corresponding position of output tensor in different reduce ops, like sum, mean, max, or min. Besides, we can use `out_size` to set necessary output shape. .. code-block:: text Given: x = [[0, 2, 3], [1, 4, 5], [2, 6, 7]] y = [1, 1, 1] src_index = [0, 1, 2, 0] dst_index = [1, 2, 1, 0] message_op = "add" reduce_op = "sum" out_size = None Then: out = [[1, 3, 4], [4, 10, 12], [2, 5, 6]] Args: x (Tensor): The input node feature tensor, and the available data type is float32, float64, int32, int64. And we support float16 in gpu version. y (Tensor): The input edge feature tensor, and the available data type is float32, float64, int32, int64. And we support float16 in gpu version. src_index (Tensor): An 1-D tensor, and the available data type is int32, int64. dst_index (Tensor): An 1-D tensor, and should have the same shape as `src_index`. The available data type is int32, int64. message_op (str, optional): Different message ops for x and e, including `add`, `sub`, `mul`, `div`. reduce_op (str, optional): Different reduce ops, including `sum`, `mean`, `max`, `min`. Default value is `sum`. out_size (int|Tensor, optional): We can set `out_size` to get necessary output shape. If not set or out_size is smaller or equal to 0, then this input will not be used. Otherwise, `out_size` should be equal with or larger than max(dst_index) + 1. Default value is `None`. name (str, optional): Name for the operation (optional, default is None). For more information, please refer to :ref:`api_guide_Name`. Returns: - out (Tensor), the output tensor, should have the same shape and same dtype as input tensor `x`. If `out_size` is set correctly, then it should have the same shape as `x` except the 0th dimension. Examples: .. code-block:: python >>> import paddle >>> x = paddle.to_tensor([[0, 2, 3], [1, 4, 5], [2, 6, 7]], dtype="float32") >>> y = paddle.to_tensor([1, 1, 1, 1], dtype="float32") >>> indexes = paddle.to_tensor([[0, 1], [1, 2], [2, 1], [0, 0]], dtype="int32") >>> src_index, dst_index = indexes[:, 0], indexes[:, 1] >>> out = paddle.geometric.send_ue_recv(x, y, src_index, dst_index, message_op="add", reduce_op="sum") >>> print(out.numpy()) [[ 1. 3. 4.] [ 4. 10. 12.] [ 2. 5. 6.]] >>> x = paddle.to_tensor([[0, 2, 3], [1, 4, 5], [2, 6, 7]], dtype="float32") >>> y = paddle.to_tensor([1, 1, 1], dtype="float32") >>> indexes = paddle.to_tensor([[0, 1], [2, 1], [0, 0]], dtype="int32") >>> src_index, dst_index = indexes[:, 0], indexes[:, 1] >>> out_size = paddle.max(dst_index) + 1 >>> out = paddle.geometric.send_ue_recv(x, y, src_index, dst_index, message_op="add", reduce_op="sum", out_size=out_size) >>> print(out.numpy()) [[ 1. 3. 4.] [ 4. 10. 12.]] >>> x = paddle.to_tensor([[0, 2, 3], [1, 4, 5], [2, 6, 7]], dtype="float32") >>> y = paddle.to_tensor([1, 1, 1], dtype="float32") >>> indexes = paddle.to_tensor([[0, 1], [2, 1], [0, 0]], dtype="int32") >>> src_index, dst_index = indexes[:, 0], indexes[:, 1] >>> out = paddle.geometric.send_ue_recv(x, y, src_index, dst_index, message_op="add", reduce_op="sum") >>> print(out.numpy()) [[ 1. 3. 4.] [ 4. 10. 12.] [ 0. 0. 0.]] r>message_op should be `add`, `sub`, `mul`, `div`, but received rr rrrr?-q=graph_send_ue_recvr"r#Yr)r*r+r,r&r'r-Tr/)r"rSr)r+) message_opr1r2r6r9) send_ue_recv)r=rr rrrUr>r r r?r@r&r'r rArr.r rBrCrrD) rEyrFrGrTr1r,rHrIrJrKr3r4s rLrUrUs7R55LZL Y  55Nyk Z   1 DAU BU 1u9 +H6JK""       OO     !  ?  !  ?  ! {$68L  ! {$68L   bhh(3$   h ) ) '"$  8vx877agg7F==> ""  %**,"*  #(   %8   rMcUS;a[SU35e[X5upUS:XaSnU*nUS:Xa SnSUS-- n[5(a&[R"XX#UR 55$[ S0[5D6n[US /S QS 5 [US /S QS 5 [US SS/S 5 [USSS/S 5 URURS9nUUUUS.nSUR 50n URS UU SU0S9 U$)a Graph Learning message passing api. This api is mainly used in Graph Learning domain, and the main purpose is to reduce intermediate memory consumption in the process of message passing. Take `x` as the source node feature tensor, take `y` as the destination node feature tensor. Then we use `src_index` and `dst_index` to gather the corresponding data, and then compute the edge features in different message_ops like `add`, `sub`, `mul`, `div`. .. code-block:: text Given: x = [[0, 2, 3], [1, 4, 5], [2, 6, 7]] y = [[0, 1, 2], [2, 3, 4], [4, 5, 6]] src_index = [0, 1, 2, 0] dst_index = [1, 2, 1, 0] message_op = "add" Then: out = [[2, 5, 7], [5, 9, 11], [4, 9, 11], [0, 3, 5]] Args: x (Tensor): The source node feature tensor, and the available data type is float32, float64, int32, int64. And we support float16 in gpu version. y (Tensor): The destination node feature tensor, and the available data type is float32, float64, int32, int64. And we support float16 in gpu version. src_index (Tensor): An 1-D tensor, and the available data type is int32, int64. dst_index (Tensor): An 1-D tensor, and should have the same shape as `src_index`. The available data type is int32, int64. message_op (str): Different message ops for x and y, including `add`, `sub`, `mul` and `div`. name (str, optional): Name for the operation (optional, default is None). For more information, please refer to :ref:`api_guide_Name`. Returns: - out (Tensor), the output tensor. Examples: .. code-block:: python >>> import paddle >>> x = paddle.to_tensor([[0, 2, 3], [1, 4, 5], [2, 6, 7]], dtype="float32") >>> y = paddle.to_tensor([[0, 1, 2], [2, 3, 4], [4, 5, 6]], dtype="float32") >>> indexes = paddle.to_tensor([[0, 1], [1, 2], [2, 1], [0, 0]], dtype="int32") >>> src_index = indexes[:, 0] >>> dst_index = indexes[:, 1] >>> out = paddle.geometric.send_uv(x, y, src_index, dst_index, message_op="add") >>> print(out.numpy()) [[ 2. 5. 7.] [ 5. 9. 11.] [ 4. 9. 11.] [ 0. 3. 5.]] rrOrrrrrPrQ graph_send_uvrE)r&r'r$r%r(rVrFr&r'rGr-)rErVrFrGrTrJ)r:r3r4r;)rX) r=rr rsend_uvr>r rBr rCr.rD) rErVrFrGrTrHrIrJr3r4s rLrYrYs]V55LZL Y   1 DAU BU 1u9 ~~aI*:J:J:LMM99  ?   !  ?   ! {Wg$6  ! {Wg$6 77agg7F""   z//12 CL   rM)rNN)rErrFrrGrr1rr,int | Tensor | NonerH str | Nonereturnr)rrNN)rErrVrrFrrGrrTrr1rr,rZrHr[r\r)rN)rErrVrrFrrGrrTrrHr[r\r)! __future__rtypingrrnumpyr@typing_extensionsrpaddlerpaddle.base.data_feederrr r paddle.base.frameworkr paddle.base.layer_helperr paddle.frameworkr utilsrrrrr__annotations__r__all__r<rUrYrMrLrjs#)' +03 " Iy $ J  !$( X XXX X " X  X X@# $(H H HH H  H  H"H H H`# ~ ~ ~~ ~  ~  ~ ~rM