Forums - caffe fast rcnn convert to dlc model issue

2 posts / 0 new
or Register
Last post
caffe fast rcnn convert to dlc model issue
zhanhe18
Join Date: 13 Sep 16
Posts: 1
Posted: Wed, 2019-07-17 07:07
Top

Hi , I have a fast-rcnn.prototxt and fast-rcnn.caffemodel and I used command :

$ snpe-caffe-to-dlc -c fast-rcnn.prototxt -b fast-rcnn.caffemodel -d fast-rcnn.dlc 

to generate dlc model.

 

This model have two input layers : data [1x3x224x224]  , im_info[1x1x1x3] ,

and three outputs which are cls_score[1x81] , bbox_pred[1x324] , cls_prob[1x81].

 

When I run this network , I can only get one output from 

zdl::DlSystem::TensorMap output_tensormap;

When I used setOutputLayers("cls_score","bbox_pred","cls_prob") it told me

error_code=204; error_message=Couldn't find name. None of the specified output layers exist!; error_component=Dl System; line_no=916; thread_id=140319638897280

and when I used setOutputTensorLayers("cls_score","bbox_pred","cls_prob") it told me
error_code=204; error_message=Couldn't find name. None of the specified output Tensors exist!; error_component=Dl System; line_no=983; thread_id=140208517300864
 
does anyone know how to get my three outputs out ? Thanks.
my snpe vesion is 1.27
  • Up0
  • Down0
Re: caffe fast rcnn convert to dlc model issue
zhengxin.zhao
Join Date: 13 Mar 19
Posts: 12
Posted: Mon, 2019-08-19 20:33
Top

zdl::DlSystem::StringList output_layer_names = {};

output_layer_names.append("name1");

output_layer_names.append("name2");

output_layer_names.append("name3");

 

have a try.

  • Up0
  • Down0
or Register

Opinions expressed in the content posted here are the personal opinions of the original authors, and do not necessarily reflect those of Qualcomm Incorporated or its subsidiaries (“Qualcomm”). The content is provided for informational purposes only and is not meant to be an endorsement or representation by Qualcomm or any other party. This site may also provide links or references to non-Qualcomm sites and resources. Qualcomm makes no representations, warranties, or other commitments whatsoever about any non-Qualcomm sites or third-party resources that may be referenced, accessible from, or linked to this site.