Deep Neural Network Python from scratch | L layer Model | No Tensorflow

We will create a Deep Neural Network python from scratch. We are not going to use Tensorflow or any built-in model to write the code, but it's entirely from scratch in python. We will code Deep Neural Network with L layers in python from scratch.

This video is for those enthusiasts who love to know under-the-hood details behind how things work. You can directly use the TensorFlow model to create a Deep Neural Network, but if you are curious to know how things work in python from scratch, then this video is for you.

Understanding Deep Neural Network in Python from scratch helps you learn how deep learning actually works and gives you confidence in understanding Machine Learning.

And if you have followed my playlist on Neural Network, then writing this code will be super simple for you. I have tried to explain a very difficult code in a simple manner, so please let me know in the comments section what you feel about this video.

➖➖➖➖➖➖➖➖➖➖➖➖➖➖➖

Timestamps:
0:00 Coming Next
0:30 Intro
3:14 Overview
6:32 Initializing Parameters
14:57 Forward Propagation
23:36 Cost Function
26:22 Backward Propagation
33:10 Update Parameters
34:23 Complete Model
40:36 Improving Model Look
48:44 End

➖➖➖➖➖➖➖➖➖➖➖➖➖➖➖

➖➖➖➖➖➖➖➖➖➖➖➖➖➖➖

➖➖➖➖➖➖➖➖➖➖➖➖➖➖➖

[mohamedkh]

Dude....your videos are FABULOUS!!
keep going!!we need you!!

[Felloggs]

Bro, your content and you as well are AWESOME.
Liked and subbed, keep it up! <3

[lokendrakumar]

Bro this is the simplest explanation in whatever I have seen

[SatyamKumar-xgsg]

just what i need to understand with clarity

[matthiasblumrich]

You and Josh Starmer (StatQuest) totally demystify DNNs. Thanks!!!

[matteomanzi]

It's a pleasure to learn with your lesson!

[lecturesfromleeds]

Straight forward and to the point! Good video

[threetime-nedc]

thank u, it the best and clear vedio I have watched😍U an extremely handsome man!!!

[engineeringdecrypted]

Thanks bro ! you are simply great teacher

[ikrameounadi]

hey, great playlists, i can now say i can really understand deep learning, can you please make a video explaining the perceptron algorithm and its complexity along with kernels ?!

[ajay]

very decent explanation, would you like to do the same for CNN?

[ashraf_isb]

Thank you so much sir, very much helpful 🙂

[svk]

Very helpful. Please make a playlist for GAN and transformers like you made for CNN.

[mustafatuncer]

I really liked it. I would also like for you to create a video about LSTMs and Transformers (from scratch).

[robertdemka]

In terms of performance, do all the string concatenations to get the parameters from dicts (the "Z" + str(l) and stuff) slow down the code, or is this trivial compared to the actual parameter multiplication?

[divyagarh]

Good stuff. No videos for a year? please keep uploading. Thank you

[soumyadeepsarkar]

thanks a lot bro your videos really helped me

[uchindamiphiri]

is this video suitable for beginers? If not recommend me what to watch before jumping to this

[ИсмоилОдинаев-йя]

Hi, Mr. Jay Patel!
Thanks a lot for such explaining!
Why you don't use derivative in the output layer (AL) for sigmoid function during backward pass?
Can we state that the weights of the last layer (WL) learn without taking into account back pass of the output error (AL-Y) through sigmoid?
If yes, why you and other guys don't use it?

[zonunmawiazadeng]

Hey I implemented Backpropagation on = dL/dA dot product on dA/dZ according to your video....In my implemented, dL/dA and dA/dZ are both the shape (training size, image height, image weight, channel size)...If this is corrected, how should we dot product it....