Deep Learning with Keras + TensorFlow - (Pt.4) Exploring Deeper CNNs

In this series we will build a CNN using Keras and TensorFlow and train it using the Fashion MNIST dataset!

In this video, we will look at 3 different CNN models, each with more layers and trainable parameters. We'll also see how these perform interms of accuracy and loss

[rong]

great job! very practical fun and straight to the point. Looking for more videos from you!

[evabenitogaragorri]

Thanks for these Mark, really helpful to have a step-by-step tutorial. Would be great to have some more theoretical background alongside the coding! :-)

[hishammohammed]

How to get the prediction for a single input image?

[josephmulingwakithome]

Hey I want to use fashion mnist in my Android that uses firebase database....am just asking for the steps that follow after training a model to get it working

[narenabbaraju]

How to print the results to a Submission.csv file with columns ImageID and Label?..please help...

[soonheng]

I am just beginner here. But when I want to point out that the validation accuracy and loss graphs (at 18:04), It seems like it is not as accurate as expected. Generally, we expect Tensorflow to perform around 0.98... of accuracy, or around 0.02...ish kind of loss to be considered acceptable (At least for me). but with 0.93~4..ish accuracy and 0.2...ish loss seems a bit too inaccurate for me.
Honestly I never tried Keras and Fashion MNIST so I am not sure if which of them are the reason. Maybe someone can shed some light here so that we can learn together?
And keep up with the videos, I learn a lot from your videos.