• State Verified Answer
  • Replies 2 replies
  • Subscribers 70 subscribers
  • Views 853 views
  • Users 0 members are here
Attachments (0)
Nordic Case Info
  • Case ID: 294244
Options

Thingy53: Support for tensorflow Lite

Divya_uni_ulm

Hello,

I have a few queries:

  • I need to confirm that does Thingy53 board support TensorFlow lite. So, for my ongoing project, I am using TensorFlow to develop my code and then I have trained the model and then converted it into tflite (basically it is a .hex file). So, is there any possibility that I can run this .hex file on Thingy53?

 

            

  • Also, I have noticed that if I want to connect my device to Edge Impulse, then according to Edge impulse review, they do not support importing pre-trained tflite model files into Edge Impulse. So, is there any chance I can use Thingy53 for my already trained model and then can proceed further?

Regards,

Divya

  • 0

    Hello Divya,

    I am looking into it, and will get back to you on this.

    Regards,

    Elfving

  • +1

    Hello Divya,

    I need to confirm that does Thingy53 board support TensorFlow lite. So, for my ongoing project, I am using TensorFlow to develop my code and then I have trained the model and then converted it into tflite (basically it is a .hex file). So, is there any possibility that I can run this .hex file on Thingy53?
    Also, I have noticed that if I want to connect my device to Edge Impulse, then according to Edge impulse review, they do not support importing pre-trained tflite model files into Edge Impulse. So, is there any chance I can use Thingy53 for my already trained model and then can proceed further?

    I see that you've already asked the Edge Impulse team about how to get pretrained tflite models into edge impulse, and that it isn't possible at the moment. 

    It might be possible to get it to work without edge impulse using TF lite micro (see here), though then you would be a bit on your own. I would recommend you to use edge impulse instead, as that is what we have support for. 

    I am confused as I have noticed that TensorFlow Lite supports only the below-mentioned boards: 

    It might be that it works on others, though that it only has built in support for or have been tested on the ones listed.

    Regards,

    Elfving

Related
Terms of service