• State TI Thinks Resolved
  • Locked Locked
  • Replies 3 replies
  • Answers 1 answer
  • Subscribers 97 subscribers
  • Views 631 views
  • Users 0 members are here
Support feedback
Options
Options
Related

This thread has been locked.

If you have a related question, please click the "Ask a related question" button in the top right corner. The newly created question will be automatically linked to this question.

Original question:

TDA2: How to run SSD based TIDL OD use case in Vision SDK with pre-trained model

TDA2: How to run pose model based TIDL SEG use case in Vision SDK with pre-trained model

user1916368
Intellectual 560 points
Part Number: TDA2

hi, dear

           1， Is there any tips about changing the seg configuration files to run user model like my pose reg model? 

           2， Beside these facts above, the output of my pose model is not the limited classification but arbitrarily float32 value. So from the link that you offer me before, I do not know if it will work successfully by just commenting the classfications.(this Line : tempoutPutList->objDesc[tempoutPutList->numObjects].objType = (label - 1); need to be changed to " (label - 1)%3 ") Any other change should be made?

            3，In fact, in my case only, if I use this case called "chanis_tidl – File IO based use case to validate user model ", it is better to bring up? If so, could you offer me some tips for changing this demo to my own pose mode recog demo?This case is similar to my pose reg demo, and I want to test the performance of the mode only, not for accuracy which maybe need only one image like in.RGB, right? And how to translate my jpg image to the in.RGB format? 
Good luck, tks!

  • 0
    TI__Genius 17580 points
    Hi,

    1. I will check and come back on this.

    2. Line : tempoutPutList->objDesc[tempoutPutList->numObjects].objType = (label - 1)%3;
    This modulo 3 is because, currently VSDK supports only four colour codes for display. This means it can display all the 21 objects but can use only 4 colours as below,
    Class 0, 4 , 8... uses first colour
    Class 1, 5, 9... uses second colour
    Class 2, 6, 10..uses third colour
    Class 3, 7, 11... uses the fourth colour coding.

    3. I recommend standalone setup (using CCS and TIDL .out file) to validate your model first, once you confirm that it is working on standalone then you can run it on VSDK use case.

    Thanks,
    Praveen
  • 0 in reply to Praveen Eppa1
    Intellectual 560 points
    Tks for sharing.
    1. The key point that I care is which file contains the line “tempoutPutList->objDesc[tempoutPutList->numObjects].objType = (label - 1)%3”？
    2. I think running on the ccs with the simulator may be useful and convenient to validate, yet I said many times, I can not get the source code of TIDL from the representative in shanghai. If I have not got that code, could I just use the limited tools for debugging on CCS? I had better get the source code for building the tools and debugging on CCS ? Right?
  • 0 in reply to user1916368
    TI__Genius 17580 points
    1. File location is shared in this thread
    e2e.ti.com/.../723749

    2. For running on the ccs with the simulator and validation, source code may not required. It may be required if you want to add new layers and debugging.

    Thanks,
    Praveen