Transfer Learning for Keras Feature Engineering (9.5)

preview_player
Показать описание
Transfer learning can be applied to feature engineering. The outputs from hidden layers inside of the pretrained neural network can be used as features. This video demonstrates how to use the hidden layer output from pretrained neural networks as features.

Code for This Video:

Follow Me/Subscribe:

Рекомендации по теме
Комментарии
Автор

From the outputs, it is possible to reverse engineer the 'features' it has discovered?

Nickle