miscellaneous ML ideas that don't really fit anywhere else.
This is an image classification pipeline which uses Hugging Face Transformers to fine-tune a 'google/vit-base-patch16-224-in21k' model on images of food. There are 101 different food classes it is trained on, and two example images are provided in food_classifier/image_files/. For more info see the model card on Hugging Face.
cd to the food_classifier directory and run python3 food_classifier.py after adding the images you want to test to food_classifier/image_files/. Results are a dictionary of image names and corresponding labels and scores