设备是:华为ATLAS 200I DK A2开发者套件,CANN版本是:7.0.rc1,使用ATC进行模型转换时出现如下报错:Expand算子不支持,但是CANN文档中是支持的,请问这是怎么回事?
报错信息如下:
ATC run failed, Please check the detail log, Try 'atc --help' for more information
EZ3002: Optype [Expand] of Ops kernel [AIcoreEngine] is unsupported. Reason: [Dynamic shape check]: data type DT_DOUBLE of input [x] is not supported. All supported data type and format of tensor input0.x is: Data Type: {DT_FLOAT16,DT_FLOAT,DT_INT32,DT_UINT8,DT_INT8,DT_BOOL,DT_FLOAT16,DT_FLOAT,DT_INT32,DT_UINT8,DT_INT8,DT_BOOL}Format:{ND,ND,ND,ND,ND,ND,ND,ND,ND,ND,ND,ND}[Static shape check]:data type DT_DOUBLE of input [x] is not supported. All supported data type and format of tensor input0.x is: Data Type: {DT_FLOAT16,DT_FLOAT,DT_INT32,DT_UINT8,DT_INT8,DT_BOOL,DT_FLOAT16,DT_FLOAT,DT_INT32,DT_UINT8,DT_INT8,DT_BOOL}Format:{ND,ND,ND,ND,ND,ND,ND,ND,ND,ND,ND,ND}[tbe-custom]:op type Expand is not found in this op store.[tbe-builtin]:[Dynamic shape check]: data type DT_DOUBLE of input [x] is not supported. All supported data type and format of tensor input0.x is: Data Type: {DT_FLOAT16,DT_FLOAT,DT_INT32,DT_UINT8,DT_INT8,DT_BOOL,DT_FLOAT16,DT_FLOAT,DT_INT32,DT_UINT8,DT_INT8,DT_BOOL}Format:{ND,ND,ND,ND,ND,ND,ND,ND,ND,ND,ND,ND}[Static shape check]:data type DT_DOUBLE of input [x] is not supported. All supported data type and format of tensor input0.x is: Data Type: {DT_FLOAT16,DT_FLOAT,DT_INT32,DT_UINT8,DT_INT8,DT_BOOL,DT_FLOAT16,DT_FLOAT,DT_INT32,DT_UINT8,DT_INT8,DT_BOOL}Format:{ND,ND,ND,ND,ND,ND,ND,ND,ND,ND,ND,ND}.
Possible Cause: The operator type is unsupported in the operator information library due to specification mismatch.
Solution: Submit an issue to request for support at https://gitee.com/ascend, or remove this type of operators from your model.
TraceBack (most recent call last):
Optype [Expand] of Ops kernel [DNN_VM_HOST_CPU_OP_STORE] is unsupported. Reason: data_type DT_DOUBLE of input[0, x] is unsupported, op type[Expand]..
Optype [Expand] of Ops kernel [aicpu_ascend_kernel] is unsupported. Reason: data_type DT_DOUBLE of input[0, x] is unsupported, op type[Expand]..
No supported Ops kernel and engine are found for [/decoder/Expand_1], optype [Expand].
No supported Ops kernel and engine are found for [/decoder/Expand_3], optype [Expand].
No supported Ops kernel and engine are found for [/decoder/Expand], optype [Expand].
No supported Ops kernel and engine are found for [/decoder/Expand_2], optype [Expand].
build graph failed, graph id:0, ret:-1[FUNC:BuildModelWithGraphId][FILE:ge_generator.cc][LINE:1615]
GenerateOfflineModel execute failed.
设备是:华为ATLAS 200I DK A2开发者套件,CANN版本是:7.0.rc1,使用ATC进行模型转换时出现如下报错:Expand算子不支持,但是CANN文档中是支持的,请问这是怎么回事?
报错信息如下:
ATC run failed, Please check the detail log, Try 'atc --help' for more information
EZ3002: Optype [Expand] of Ops kernel [AIcoreEngine] is unsupported. Reason: [Dynamic shape check]: data type DT_DOUBLE of input [x] is not supported. All supported data type and format of tensor input0.x is: Data Type: {DT_FLOAT16,DT_FLOAT,DT_INT32,DT_UINT8,DT_INT8,DT_BOOL,DT_FLOAT16,DT_FLOAT,DT_INT32,DT_UINT8,DT_INT8,DT_BOOL}Format:{ND,ND,ND,ND,ND,ND,ND,ND,ND,ND,ND,ND}[Static shape check]:data type DT_DOUBLE of input [x] is not supported. All supported data type and format of tensor input0.x is: Data Type: {DT_FLOAT16,DT_FLOAT,DT_INT32,DT_UINT8,DT_INT8,DT_BOOL,DT_FLOAT16,DT_FLOAT,DT_INT32,DT_UINT8,DT_INT8,DT_BOOL}Format:{ND,ND,ND,ND,ND,ND,ND,ND,ND,ND,ND,ND}[tbe-custom]:op type Expand is not found in this op store.[tbe-builtin]:[Dynamic shape check]: data type DT_DOUBLE of input [x] is not supported. All supported data type and format of tensor input0.x is: Data Type: {DT_FLOAT16,DT_FLOAT,DT_INT32,DT_UINT8,DT_INT8,DT_BOOL,DT_FLOAT16,DT_FLOAT,DT_INT32,DT_UINT8,DT_INT8,DT_BOOL}Format:{ND,ND,ND,ND,ND,ND,ND,ND,ND,ND,ND,ND}[Static shape check]:data type DT_DOUBLE of input [x] is not supported. All supported data type and format of tensor input0.x is: Data Type: {DT_FLOAT16,DT_FLOAT,DT_INT32,DT_UINT8,DT_INT8,DT_BOOL,DT_FLOAT16,DT_FLOAT,DT_INT32,DT_UINT8,DT_INT8,DT_BOOL}Format:{ND,ND,ND,ND,ND,ND,ND,ND,ND,ND,ND,ND}.
Possible Cause: The operator type is unsupported in the operator information library due to specification mismatch.
Solution: Submit an issue to request for support at https://gitee.com/ascend, or remove this type of operators from your model.
TraceBack (most recent call last):
Optype [Expand] of Ops kernel [DNN_VM_HOST_CPU_OP_STORE] is unsupported. Reason: data_type DT_DOUBLE of input[0, x] is unsupported, op type[Expand]..
Optype [Expand] of Ops kernel [aicpu_ascend_kernel] is unsupported. Reason: data_type DT_DOUBLE of input[0, x] is unsupported, op type[Expand]..
No supported Ops kernel and engine are found for [/decoder/Expand_1], optype [Expand].
No supported Ops kernel and engine are found for [/decoder/Expand_3], optype [Expand].
No supported Ops kernel and engine are found for [/decoder/Expand], optype [Expand].
No supported Ops kernel and engine are found for [/decoder/Expand_2], optype [Expand].
build graph failed, graph id:0, ret:-1[FUNC:BuildModelWithGraphId][FILE:ge_generator.cc][LINE:1615]
GenerateOfflineModel execute failed.