Hello

First of all, great repo! I particularly want to thank you for your help in creating export_tfv2_lite_models.ipynb, I was able to convert my custom model with SSD MobileNet V2 FPNLite 640x640 to edge_tpu.tflite smoothly!

However, I am having difficulty understanding your repo to "run the model on a raspberry pi with coral"..

I have followed your guide from https://github.com/NobuoTsukamoto/edge_tpu/tree/master/detection/cpp

But I think I am running C++ instead of object_detection_capture_picamera.py?

I was wondering if you can provide me a guide on what I need to do in order to run your object_detection_capture_picamera.py script on my Raspberry Pi 4 2GB with Coral.

I have tried installing some of the dependencies myself but I still have no luck in running the script to see my model live on my Pi. I have cloned your entire repo but I don't think it is needed for what I want, maybe I am wrong?

Any help at all you can that you can share in order to run the correct repo/script on my Pi will be very appreciated.

Thank you again for your work!

Hi, amazing repo, I got my hands on an Edge TPU but I am struggling in finding an actually useful way to use it. I am interested in doing only instance segmentation (so no classification whatsoever), and I was wondering if in your semantic segmentation approach based on DeepLabV3 you are limited to the same 20 classes listed here.

Thanks

Here. :smiley: https://github.com/NobuoTsukamoto/tflite-cv-example/tree/master/deeplab_edgetpu2#coral-dev-board-edgetpu-autosegedgetpu-xs-with-fusing-argmax