Previously we were using images of single parts as training data but now we are going to try using images of multiple parts bunched together.

I'm still researching this but we have decided that we need a really small model as it will need to be trained on the spot

the main feature is the ability to take a picture of any box or bin and it will identify and count all of the items inside of it. I have more stuff planned but for now im just trying to get this feature working