Forums - onnx and dlc outputs are not matching

2 posts / 0 new
Last post
onnx and dlc outputs are not matching
arushi_arora
Join Date: 21 Mar 23
Posts: 1
Posted: Wed, 2023-05-24 05:14

Convert pytorch model to onnx model using this link: https://github.com/fire717/movenet.pytorch/tree/6eb59cd47bfde753530ed236...

then converted onnx to dlc using snpe 2.9

torch version: 2.0.0

onnx version:1.6.0

numpy v1.16.5

  • protobuf v3.6.0

 

datatype: float32

When comparing output generated using onnx model and output generated using dlc model, getting huge difference.

Does anyone else faced same issue? Is there any way to debug and resolve it?

 

  • Up0
  • Down0
weihuan
Join Date: 12 Apr 20
Posts: 270
Posted: Fri, 2023-05-26 22:01

Dear developer,

What's backend in your application?

You also could compare the results generated by SNPE-CPU to check if any accuracy happened.

BR.

Wei

  • 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.