Deep Learning - Choosing Network Size

preview_player
Показать описание

How many nodes and layers do we need? We combine elements of scikit learn and Keras Neural Nets in this lesson.

Рекомендации по теме
Комментарии
Автор

Thanks, can you help me to understand keras tuner optimization

الدينوالحياة-خي
Автор

You do very useful videos. I learn a lot. Could you please explain from where I start for LSTM initial layer size? Next Layer size? Depends in quantity of inputs/features?

whocareswhoiamday
Автор

Sir is there an explanation for why we go for the pyramidal structure?

shahbazquraishy
Автор

I would have thought that the bigger the batch size the better, does any body have an explanation for this result?

sergior.m.
Автор

I can't get my collab to read the csv

timmik.rysgaard
Автор

The github link is broken. It would be help if you can you provide the alternate link.

KrishnanandSinghGeek
Автор

interesting but i wish the option set was pre-defined

MrSupportDemocracy
Автор

There is sound but it seems to not always be there. Otherwise very good. Thanks

slideroolz
Автор

It's really helpful. But how did you selected the number 30 for layers? is it dependent on number of input features or you selected it randomly?

saumyashah
Автор

I guess the grid_result.best_score_ is the test score rather than the train score. There is no validation_split in the .fit() call, so how is the test data taken?

rajilsaraswat