Forums - Relationship for between "Android Neural Network API" and "SNPE"

6 posts / 0 new
Last post
Relationship for between "Android Neural Network API" and "SNPE"
Ryosuke MATSUZAKA
Join Date: 21 Aug 17
Posts: 1
Posted: Sun, 2018-11-25 22:21
From Android O (v8.1), Neural networks API provides accelerated computation and inference for on-device machine learning frameworks by Google.
 
Is it possible to use SNPE by using this Google NN API ?
This means that we are considering whether Google NN API and SNPE can be integrated or not.
 
Or, we would like to run edge computing on DSP(Hexagon) or GPU by using NN API.
Is it possible?
 
BR,
 

 

  • Up0
  • Down0
rafalfirlejczyk
Join Date: 19 Oct 18
Posts: 3
Posted: Tue, 2018-12-18 12:38

I think it is not possible to use Google Android NN APIs for the Snapdragons, because most of the Snapdragons does not support the Google Android NN API. According to my understanding the support for Android’s NN will appear first in Snapdragon 845. QC staff - please correct me, if I am wrong.

  • Up0
  • Down0
chenfeng
Join Date: 30 Aug 17
Posts: 6
Posted: Thu, 2018-12-20 23:23

Hi,

SNPE JAVA API is difference with Google NN API.   So current can not integrated together.
But as I knew, QCT will support Android NN API.  Once ready, I think it will disclose it.

 

Thanks.

  • Up0
  • Down0
rafalfirlejczyk
Join Date: 19 Oct 18
Posts: 3
Posted: Tue, 2019-03-19 05:14

Does Qualcomm intend to support the Google NN API on SD660?

  • Up0
  • Down0
rafalfirlejczyk
Join Date: 19 Oct 18
Posts: 3
Posted: Tue, 2019-04-02 01:55

I do not think this forum is read or mantained by Qualcomm - no answer since days, weeks...

  • Up0
  • Down0
gesqdn-forum
Join Date: 4 Nov 18
Posts: 184
Posted: Wed, 2019-05-15 02:30

Hi,
The SNPE SDK doesn't have the Google NN API's integrated.

The SNPE SDK consists of,

  1. Java library for Android integration
  2. C++ and python library to help in user implementation to modify the usage of the machine learning model
  3. Tools that convert the network models to .dlc format
  4. Benchmarking and Throughput calculation tools
  • 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.