A model that recognises whether there's fluffy things in an image.
The model had a 98% accuracy on the validation set. Here's a couple of accurate predictions:
I also created 2 Bash scripts:
- img-prep.sh - prepares the images for use in the notebook after they've been downloaded
- revert.sh - replicates how I have the raw, unarchived images in my local environment based on the fluffy-images archive; this was useful when I accidentally deleted the raw images 😅