Food 101 challenge summary
Food-101 is a challenging vision problem, but everyone can relate to it. Recent SoTA is ~80% top-1, 90% top-5. These approaches rely on lots of TTA, large networks and even novel architectures.
Train a decent model >85% accuracy for top-1 for the test set, using a ResNet50 or smaller network with a reasonable set of augmentations.
Original Kaggle kernel is here: https://www.kaggle.com/phananhvu/fellowship-ai-food-101-fast-ai-v6
GitHub may fail to load the notebook. In that case, the notebook can be viewed here: https://nbviewer.jupyter.org/github/phanav/food-101/blob/master/fellowship-ai-food-101-fast-ai-v6.ipynb