A machine learning process utilizing supervised learning and discriminative techniques to recognize a living being, a landscape or any other thing in images
- images per animals : 48 for train 12 for validation (some doubles)
- batch_size : 20
- images per animals : 200 for train, 50 for validation (some doubles and some bad images)
- batch_size : 4
- new model extension : .keras
- batch_size = 16
- greatly improved code organization, readability, and documentation
- transformation of the project to make it broader: moving from animals to entities
- the scraper is functional from A to Z
- ./models/ : versions of the AI
- ./data/ : statistical version tracking
- ./entities.txt : the list of entities, such as an animal or an object
- ...
✔️ : new existing model
✖️ : no new model exists