Recognize fruits! Because, why not?
It's currently using the Fruits-360 dataset
The models I've been trying are doing great validation and training accuracy wise, however, when it comes to extracing features it's getting a little confused, as it'll constantly mistake avocado for Cucumber with 90%+ confidence, my thought is that it's extracting the "texture" as a feature, but not taking into account overral shape and color.
/models- All trained models are saved here with some information about them on the title.main.py- Holds the current code for training the modelpredicter.py- Run this to predict results on a given image or generatorFigures_{}.png- Val/Train Loss and Accuracy.
Results from running a model, this is the average accuracy and loss.
