• State TI Thinks Resolved
  • Locked Locked
  • Replies 3 replies
  • Answers 1 answer
  • Subscribers 99 subscribers
  • Views 96 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.

TDA4VH-Q1: TIDL: Custom new model guideline: Issue with generating artifacts

Khoi Nguyen
Prodigy 50 points
Part Number: TDA4VH-Q1

Tool/software:

Im currently working with the custom model follow this guideline edgeai-tidl-tools/docs/custom_model_evaluation.md at master · TexasInstruments/edgeai-tidl-tools. When i add a new models config in model_configs.py, and build it by this command:
python3 ./tflrt_delegate.py -d -m cl-tfl-custom
Nothing shows up at my custom artifacts model. Yea and my models is .tflite extension, it was trained with only 1 layer. Did i do wrong or miss something??

  • 0
    TI__Genius 14420 points

    Hi Khoi,

    In the command line you are using -d, this means it is running in host mode, so no artifacts are produced.  If you want artifacts, use -c -m.  Also why are you using TF?  The industry is settling on ONNX as the common interchange format.  It looks like you are starting on this project and if you can switch I would recommend ONNX model format.

    Regards,

    Chris

  • 0 in reply to Chris Tsongas
    Prodigy 50 points

    Hi Chris, 
    I'd already used the -c and -m flag. But the result is the same. I still dunno why my folder only have artifacts and models but nothing inside at all. And yea, today i'll try to use the ONNX format, if they could do the diff. I will inform u later!
    Best,
    Khoi

  • 0 in reply to Khoi Nguyen
    TI__Expert 4430 points

    The engineer responsible is currently out of office. Please expect delay in response until next week.

    We appreciate your patience.

    Warm regards,

    Christina