How to train a custom TFLITE model? (RPi4B)

Hi! I’m using a Raspberry pi 4 B to detect phones. I want to train my own model using my family’s and friends’ phones and I got a database to do that of 50 photos from different angles.

Now… How can I train my own .TFLITE model, and most importantly, how can I make it work? (would appreciate a python code as well) I used this tutorial so far and everything worked fine, but I’m willing to change the code if it means I can finally do a detection

I’m a beginner in this, so do let me know if I’m missing on something and it was an obvious answer

TensorFlow Lite Model Maker simplifies the process of training a TensorFlow Lite model using custom dataset. It uses transfer learning to reduce the amount of training data required and shorten the training time.

The Model Maker library currently supports image classification, Object Detection, Text Classification, BERT Question Answer, Audio Classification etc.

So you can use the TensorFlow Lite Model Maker for training lite model using custom dataset and run inference using tflite_runtime for your use case.

Thank you!