generate text from photos of tree bark, matched against a corpus using embeddings.

Code: https://github.com/martin-martin/barkprints