CNN based model trained to classify 131 categories of fruits and vegetables.
CNN was handcrafted and hyer-parameters were appropriately tuned. Furthermore, Data Augmentation and Dropout were used to reduce over-fitting. This model is trained on the fruits-360 dataset available at the following link: https://www.kaggle.com/moltean/fruits
Model achives 99% accuracy on train set and 98.7% accuracy on test set in less than 10 epochs.
Model weights can be downloaded from: https://cmu.box.com/s/8qwgpwhqdcii0xrd31r2atquypcpw4us