Introduction to Keras Transfer Learning (9.1)

preview_player
Показать описание
Transfer learning is a process that loads weights from previously trained neural networks. This video shows you how to train a neural network and transfer its weights to another.

Code for This Video:

Follow Me/Subscribe:

Рекомендации по теме
Комментарии
Автор

This video is absolutely gold! I've said it before and I'll say it again, your channel and the information you put out for free is top notch!

leonardsmith
Автор

Jeff you are great thank you.
I have one question or misunderstanding in mind, could you please help me out?
How can you fit the second Data to a NN which only has non trainable parameters? It was like 1500 non trainable neurons. Then you only add the final output layer with one neuron. How and why does This work?
Add on question: would it be wise to add more layers to the pretrained layers that can be trained? Instead of just adding the final output layer?

patrickes
Автор

Transformers in the new GPU's sound interesting and maybe more on that and when it comes to the desktop. Will they replace CNN's? Also when is the drawing for the GPU? Daddy needs a new RTX with some memory!

davidwhite
Автор

no drunken sales man? U know, the 1 that fights the bull. sorry! reminds me of logic nodes? thanks 4 laying it out!

__--JY-Moe--__