Hemanth-Jonnala

Hello @abucka I think, it is possible. Please feel free to try to retrain pure tflite models from [test_data](https://github.com/google-coral/test_data) repo using back propagation method.

The process might be hanging due to CPU limitation. Can you try with one or two images from each dateset?

Hello @iverona this project needs to be updated with [pycoral API](https://github.com/google-coral/pycoral) as [edgeTPU API](https://github.com/google-coral/edgetpu) is deprecated. Unfortunately, we don't have any planned updates for this project.

@Xusniyor we don't have support for virtual machines. On windows, please try the [Arduino IDE 2.0.4](https://github.com/arduino/arduino-ide/releases/tag/2.0.4) to flash the coral micro examples. Thanks!!

On dev board please try the below steps to connect to [dev board micro](https://coral.ai/docs/dev-board-micro/serial-console/#connect-over-usb-otg). Thanks!! ``` sudo apt-get install screen sudo usermod -aG plugdev,dialout $USER sudo reboot now screen /dev/ttyACM0...

please refer to this comment: https://github.com/google-coral/coralmicro-out-of-tree-sample/issues/6#issuecomment-1515293630

Please check the input data type for TFLITE_detection_post_process op. It should be float32..Thanks!! 

I am not sure whether TF1.x models would work on Dev Board Micro. As there is no support for TF1.x on colab now, I suggest trying to generate models with...

Could you share the script that you are trying to run here please.Thanks!

Yes, please share all the reproducable steps, files etc. - And are you planning to run this model on any microcontrollers? - Have you tested any other [lower resolution models](https://coral.ai/models/pose-estimation/#trained-models)?