Learning to use the fast.ai toolchain, I proposed this excersise to myself to practice both sourcing data from api calls to image search providers and to practice using the fast.ai tools.
This is an attempt to make a classifier which can detect if an image is of a beercan, bacon or neither. The classifier is a simple homage to the awkward time of the internet between web1.0 and web2.0, where a single serving humour website brought to the attention that beercan and bacon sound identical in certain accents.
Now uploaded to binder at: https://mybinder.org/v2/gh/Pandalism/beercan_or_bacon/main?urlpath=%2Fvoila%2Frender%2Finfer_beer_or_bacon.ipynb