The documentation refers to RPI OS bullseye. When can we expect support for RPI OS bookworm (first release 2023-10-10)?
1 Like
yesterday I made a machine learning model using edge impulse for image classification, but I did the machine learning operation using a raspberry pi 4B with Bookworm OS with usb camera, because I don’t have much time to read documentation (need to finish the project quickly) (I’ll learn later ) machine learning inference using edge impulse for linux ( .eim), plus I needed a lot of customization, I decided to use the inference method using tflite-runtime, and it worked. here’s what I did:
Take the model that has been trained (on the dashboard), select the .H5 extension, then convert the model using a converter to .tflite, (I use Google Colab). why not .lite? I haven’t studied it much, but tflite-runtime works on .tflite not .lite, although it may be similar.
Then I created a machine learning inference program for tflite-runtime, and that was it. maybe you need a lot of customization for your program.
happy coding 