On device training of TFLite

General Discussion
,
#1

Can I add more personalization layers to on device training of TFlite, it seems that only the head layer can be added in the case?

#3

I’m very sorry, it was my mistake, I’ve figured out how to add more layers as well as train it flexibly (defining the relevant interpreter functions on the server), thank you!

#5

I am asking myself if it’s possible to change the last layer (or any layer) of my TFLite model on the device. I know that updating weights is possible. But let’s say, I have a classification problem with a Dense layer in the end giving me 10 outputs, and I want to add a new class. This means I want to extend my Dense layer from 10 to 11. Is that possible? And if so, how? Thanks in advance!

#7

Hi @ konak

You have mentioned On device weight updation of tflite model is possible. Can you please share some resources regarding On-device weight updation in IOT devices?

Thanks!