My max accuracy is 72%

my max accuracy is 72%
with layer = [10,7,5,3]
epochs = 500,
learning_rate = 0.00001

what should i change to improve my accuracy?

hey @itzsakshi26gupta ,
can you please provide link to your code so that i am able to understand your approach.

Thank You and Happy Learning :slightly_smiling_face:.

https://drive.google.com/file/d/1rcGDIkJrt9bS_6k4OGsXHx_RjpHHxYHy/view?usp=sharing

hey ,
you have provided me .html file. Kindly check it.

https://drive.google.com/file/d/1hXOPWP8l35n3ZD72SgS0c5vFkJUynMnr/view?usp=sharing

hey @itzsakshi26gupta ,
i saw your code. Its really good.
With currently working from scratch on neural networks we are currently restricted to use only softmax and tanh activation functions , where as there are many other too.
Currently you can only alter learning rate or number of iterations, i did many experiments and found the layers [ 5,4 ] produces good results.
So, if you want to learn neural network from scratch then the code you have done is extremely good in respect of it , but if you need to increase your accuracy then you go for deep learning frameworks like tensorflow or keras , they can help you achieve much much better score.

I hope this helps you.
Thank You and Happy Learning :slightly_smiling_face:.

layers [ 5,4 ] produces good results
gave me 25 % accuracy, but anyways I will improve my score by trying different algorithms.
Thank You :slight_smile:

what validation score did get on them .
I was getting 0.6931

That’s why i said you this.

i got 0.24 something

Okay , there might be some mistake in my code.