LICENSE: https://github.com/burliEnterprises/tensorflow-image-classifier
http://www.talkapps.org/bulk-image-downloader
https://codelabs.developers.google.com/codelabs/tensorflow-for-poets/#0
https://ai.googleblog.com/2017/06/mobilenets-open-source-models-for.html
Download the builk-image-downloader. Then google and search for bill gates, steve jobs, markzuckerberg in the images tab, choose the tools => type => face, at this point click the bulk image downloader, select current tab, then it is will select all, so first click invert selection, then select minimum 20 pages, that is only the given person and it will be in the Downloads folder and it will be like image.jpg, image(1).jpg, select these and save to tensorflow-images-classifier/training_dataset/name-of-person.
So you should have all 3 people at least each 20 pictures. Now let's train.
https://www.tensorflow.org/hub/tutorials/image_retraining
sudo apt install python3-setuptools python3-pip
pip3 install tensorflow tensorflow_hub
curl -LJO https://raw.githubusercontent.com/googlecodelabs/tensorflow-for-poets-2/master/scripts/retrain.py
./train.sh
./classify.py validate-images/gates1.jpg
./classify.py validate-images/gates2.jpg
./classify.py validate-images/gates3.jpg
./classify.py validate-images/gates4.jpg
./classify.py validate-images/gates5.jpg
./classify.py validate-images/gates6.jpg
./classify.py validate-images/gates7.jpg
./classify.py validate-images/gates8.jpg
./classify.py validate-images/zuker1.jpg
./classify.py validate-images/zuker2.jpg
./classify.py validate-images/zuker3.jpg
./classify.py validate-images/zuker4.jpg
./classify.py validate-images/zuker5.jpg
./classify.py validate-images/zuker6.jpg
./classify.py validate-images/zuker7.jpg
./classify.py validate-images/zuker8.jpg