Thanks to @mootari for a suggestion that got me over the hump with this notebook:
Use Google’s Teachable Machine to:
- Gather images for an image identification model;
- Train your model
- Export and Import(via file upload) into Observable and play away.