The Deep learning model is built using fast.ai library and kaggle dataset called fruits-360
Total number of images: 82213.
Training set size: 61488 images (one fruit or vegetable per image).
Test set size: 20622 images (one fruit or vegetable per image).
Multi-fruits set size: 103 images (more than one fruit (or fruit class) per image)
Number of classes: 120 (fruits and vegetables).
Image size: 100x100 pixels.
The whole dataset is trained on resnet34 architecture and with which I managed get an accuracy of 99.9%
