Brilliant, I never thought the DL can be tough so clear and understandable
I have some questions here:
- Image size: If my images resolution are in different size, 600X800, 1280X720…etc. How could I prepare those training data-set?
- Mode-Save: If I want to build a mode that contains birds, cats, dogs image recognition as model.save(“bird-cat-dog classifier.tfl”). How could I start to do it?
- Image similarity: If I have trained a mode called bird-classifier.tfl, I input one dog image into classifier, then can it tells me the image similarity percentage? For example, the image is 15% similarity to the bird, not telling only True or False.
Thank you very much for my questions.