12 June, 2021 | By Magnus Frejd |
To model this information, we utilized two approaches:
3-Layer Model: i did not expect the 3 layer model to do well. Whenever we develop any model, my objective is to get a model that is dumb first. This is my stupid model. We utilized a really architecture that is basic
The ensuing accuracy ended up being about 67%.
Transfer Learning making use of VGG19: The difficulty utilizing the 3-Layer model, is i am training the cNN on an excellent little dataset: 3000 pictures. The greatest cNN that is performing train on scores of pictures.
Being a total result, we utilized a method called “Transfer Learning.” Transfer learning, is simply going for a model some other person built and utilizing it on your own data that are own.