After model testing, it works fairly well... with bottles. Due to the plurality of sample data being images of bottles, test samples with similar materials but different shapes seemed to reduce performance of model.
For the time being, I want to focus on a separate model which can identify the material of the bottle with some extra image data added to the sample. This should improve performance with other objects and add some functionality to the application.
Log in or sign up for Devpost to join the conversation.