--- title: 单算子描述文件配置 description: "不同输入或者不同Format场景，单算子描述文件配置不同，本章节给出各场景的配置示例。" url: https://www.hiascend.com/document/detail/zh/canncommercial/latest/devaids/atctool/atlasatc_16_0034.html sourcePath: /source/zh/canncommercial/900/devaids/atctool/atlasatc_16_0034.html indexId: 48ae023fb6807383d90b1f819ca33e1bec9953db3f4784cb4c383c003c83b02d67 --- # 单算子描述文件配置 不同输入或者不同Format场景，单算子描述文件配置不同，本章节给出各场景的配置示例。 本章节中的单算子是基于Ascend IR定义的，描述文件为JSON格式。关于JSON描述文件中各参数的解释请参见表1，关于单算子的Ascend IR定义请参见“Ascend IR算子规格说明(https://www.hiascend.comdocument/detail/zh/canncommercial/900/API/aolapi/operatorlist_00093.html)”。 - Format为ND： 该示例中的单算子转换后的离线模型为：add.om 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 [ { "op": "Add", "name": "add", "input_desc": [ { "format": "ND", "shape": [3,3], "type": "int32" }, { "format": "ND", "shape": [3,3], "type": "int32" } ], "output_desc": [ { "format": "ND", "shape": [3,3], "type": "int32" } ] } ] - Format为NCHW： 该示例中的单算子转换后的离线模型为：conv2d.om 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 [ { "op": "Conv2D", "name": "conv2d", "input_desc": [ { "format": "NCHW", "shape": [1, 3, 16, 16], "type": "float16" }, { "format": "NCHW", "shape": [3, 3, 3, 3], "type": "float16" } ], "output_desc": [ { "format": "NCHW", "shape": [1, 3, 16, 16], "type": "float16" } ], "attr": [ { "name": "strides", "type": "list_int", "value": [1, 1, 1, 1] }, { "name": "pads", "type": "list_int", "value": [1, 1, 1, 1] }, { "name": "dilations", "type": "list_int", "value": [1, 1, 1, 1] } ] } ] - Tensor计算过程中使用的Format与原始Format不同 ATC模型转换时，会将origin_format与origin_shape转成离线模型需要的format与shape。 该示例中的单算子转换后的离线模型为：add.om 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 [ { "op": "Add", "name": "add", "input_desc": [ { "format": "NC1HWC0", "origin_format": "NCHW", "shape": [8, 1, 16, 4, 16], "origin_shape": [8, 16, 16, 4], "type": "float16" }, { "format": "NC1HWC0", "origin_format": "NCHW", "shape": [8, 1, 16, 4, 16], "origin_shape": [8, 16, 16, 4], "type": "float16" } ], "output_desc": [ { "format": "NC1HWC0", "origin_format": "NCHW", "shape": [8, 1, 16, 4, 16], "origin_shape": [8, 16, 16, 4], "type": "float16" } ] } ] - 输入指定为常量 该场景下，支持设置为常量的输入，新增is_const和const_value两个参数，分别表示是否为常量以及常量取值，const_value当前仅支持一维list配置，具体配置个数由shape取值决定，例如，如下样例中shape为2，则const_value中列表个数为2；const_value中取值类型由type决定，假设type取值为float16，则单算子编译时会自动将const_value中的取值转换为float16格式的取值。 该示例中的单算子转换后的离线模型为：resizeBilinearV2.om 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 [ { "op": "ResizeBilinearV2", "name": "resizeBilinearV2", "input_desc": [ { "format": "NHWC", "name": "x", "shape": [ 4, 16, 16, 16 ], "type": "float16" }, { "format": "NHWC", "is_const": true, "const_value": [49, 49], "name": "size", "shape": [ 2 ], "type": "int32" } ], "output_desc": [ { "format": "NHWC", "name": "y", "shape": [ 4, 48, 48, 16 ], "type": "float" } ], "attr": [ { "name": "align_corners", "type": "bool", "value": false }, { "name": "half_pixel_centers", "type": "bool", "value": false } ] } ] - 可选输入（optional input）： 当存在可选输入，且可选输入没有输入数据时，则必须将可选输入的format配置为RESERVED，同时将type配置为UNDEFINED；若可选输入有输入数据时，则按其输入数据的format、type配置即可。 该示例中的单算子转换后的离线模型为：matMulV2.om 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 [ { "op": "MatMulV2", "name": "matMulV2", "input_desc": [ { "format": "ND", "shape": [16, 16], "type": "float" }, { "format": "ND", "shape": [16, 16], "type": "float" }, { "format": "RESERVED", "shape": [], "type": "UNDEFINED" }, { "format": "RESERVED", "shape": [], "type": "UNDEFINED" } ], "attr": [ { "name": "transpose_x1", "type": "bool", "value": false }, { "name": "transpose_x2", "type": "bool", "value": false } ], "output_desc": [ { "format": "ND", "shape": [16, 16], "type": "float" } ] } ] - 输入个数不确定（动态输入场景）： 该场景下，单算子的输入个数不确定。此处以AddN单算子为例。该示例中的单算子转换后的离线模型为：addN.om - 构造的单算子JSON文件使用动态输入dynamic_input参数，而不使用Tensor的名称name参数。 该场景下算子的dynamic_input取值必须和算子信息库中该算子定义的输入name的取值相同。具体设置几个输入，由AddN单算子描述文件属性参数中N的取值决定，用户可以自行修改输入的个数，但是必须和属性中N的取值匹配。（该说明仅针对AddN算子生效，其他动态输入算子的约束以具体算子为准。） 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 [ { "op": "AddN", "name": "addN", "input_desc": [ { "dynamic_input": "x", "format": "NCHW", "shape": [1,3,166,166], "type": "float32" }, { "dynamic_input": "x", "format": "NCHW", "shape": [1,3,166,166], "type": "int32" }, { "dynamic_input": "x", "format": "NCHW", "shape": [1,3,166,166], "type": "float32" } ], "output_desc": [ { "format": "NCHW", "shape": [1,3,166,166], "type": "float32" } ], "attr": [ { "name": "N", "type": "int", "value": 3 } ] } ] - 构造的单算子JSON文件使用Tensor的名称name参数，而不使用动态输入dynamic_input参数。 该场景下算子的name取值必须和算子原型定义中算子的输入名称相同，根据输入的个数自动生成x0、x1、x2……。具体设置几个Tensor名称，由AddN单算子描述文件属性参数中N的取值决定，用户可以自行修改Tensor名称的个数，但是必须和属性中N的取值匹配，例如N取值为3，则name取值分别设置为x0、x1、x2。（该说明仅针对AddN算子生效，其他动态输入算子的约束以具体算子为准。） 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 [ { "op": "AddN", "name": "addN", "input_desc": [ { "name":"x0", "format": "NCHW", "shape": [1,3,166,166], "type": "float32" }, { "name":"x1", "format": "NCHW", "shape": [1,3,166,166], "type": "int32" }, { "name":"x2", "format": "NCHW", "shape": [1,3,166,166], "type": "float32", } ], "output_desc": [ { "format": "NCHW", "shape": [1,3,166,166], "type": "float32" } ], "attr": [ { "name": "N", "type": "int", "value": 3 } ] } ]