Lot of confusion in this video because Prateek sir explained everything at very high pace, pls help me to understand this, or suggest some other recourse to clarify my doubts
In Video Building Fashion MNIST
hey @lgoyal50_be19,
can you be please let me know what problems are facing one by one , so that i can help you in resolving them.
Thank You 
.
What is difference between Validation, Train Accuracy and loss??
Validation Acc vs Train Acc
Validation loss vs Train loss
Assume we have dataset of 10000 records. ( full data )
So just examine the working of our model , we first split it into 2 subsets training and testing.
Now , we again split training into another two sets training and validation.
Now , we will actually train our model on training data, means our model will learn things from this model.
and while we are training our model , then at each epoch we test it on validation data , to just check that it doesn’t get overfitted and we can also examine or understand that how much performance is getting better or not.
Once , our training is completed , then we test our model on testing data, a complete unseen data .
So that we can be more sure , how it is going to perform or real world unseen data.
I hope you might have a overview from this.
So validation is just a subset of training data , to just get an expected value of the model on unseen data.
I hope I’ve cleared your doubt. I ask you to please rate your experience here
Your feedback is very important. It helps us improve our platform and hence provide you
the learning experience you deserve.
On the off chance, you still have some questions or not find the answers satisfactory, you may reopen
the doubt.