Pb between c++ and python

[BenLag2906]

Hi,

I have a problem my train is done in python, and when i try to use my json file and params in c++.

It crashed at MXPredSetInput althought it's ok with other Json file like SSD.

My Json file :

"nodes": [
{
"op": "null",
"name": "data",
"inputs": []
},
{
"op": "_copy",
"name": "id",
"inputs": [[0, 0, 0]]
},
{
"op": "null",
"name": "bn_data_gamma",
"attr": {
"eps": "2e-05",
"fix_gamma": "True",
"momentum": "0.9"
},
"inputs": []
},

Could you please give a piece of advise ?

Regards

[ThomasDelteil]

@BenLag2906 it looks like your network didn't export properly, can you put the code of how you exported this network ? Thanks!

[lupesko]

@indhub can you please remove the label "Question" and add "Bug"? this seems like a bug report, not a question.

[indhub]

@lupesko There isn't enough information to consider this a bug. That JSON file doesn't look normal. I would wait for the user to provide more information about the model and how it was exported.

[BenLag2906]

Hi,
I give you my entire json file . In c++ it crashed, in python it's OK to launch CNN.

yks_pose-symbol.zip

regards

`

[BenLag2906]

Hi,
For adding a precision, I use the same deploy script as SSD. Is there another way deploy it ?

regards.

[BenLag2906]

To add more explanation, i try to translate kohillyang code in c++, but many difficulties persist.

[leleamol]

@mxnet-label-bot update [Question, C API]