atc模型转换问题
收藏回复举报
atc模型转换问题
t('forum.solved') 已解决
发表于2024-07-13 20:07:35
0 查看

我有一个onnx模型,

输入为

input:

name: fog_image

tensor: float32[batch,3,height,width]

name: omega

tensor: float32

output:

name: clear_image

tensor: float32[batch,3,height,width]

请问如何写atc指令
atc --model=dcp.onnx --framework=5  --soc_version=Ascend310B1 --input_shape='fog_image:1,3,255,255;omega:1' --output=dcp
报错:
ATC run failed, Please check the detail log, Try 'atc --help' for more information 

E89999: Inner Error! 

E89999  op[/Concat_1], dtype of inputs must be same, x2:DT_FLOAT, x0:DT_INT64.[FUNC:CheckInputDtypeSame][FILE:util.cc][LINE:133] 

        TraceBack (most recent call last): 

        Verifying /Concat_1 failed.[FUNC:InferShapeAndType][FILE:infershape_pass.cc][LINE:137] 

        Call InferShapeAndType for node:/Concat_1(ConcatD) failed[FUNC:Infer][FILE:infershape_pass.cc][LINE:119] 

        process pass InferShapePass on node:/Concat_1 failed, ret:4294967295[FUNC:RunPassesOnNode][FILE:base_pass.cc][LINE:530] 

        build graph failed, graph id:0, ret:1343242270[FUNC:BuildModel][FILE:ge_generator.cc][LINE:1438] 

我要发帖子