0
$\begingroup$

When training neural network, is it better to randomly choose data for every batch or feed the data sequentially? Why?

Improve this question
$\endgroup$

1 Answer 1

Reset to default
1
$\begingroup$

If you feed the data randomly, there are fewer chances of overfitting than if you feed the data sequentially.

If the data is fed sequentially, then the network will not learn to generalize as the weights will be updated for that batch according to that same image again and again.

It is always recommended to feed the randomly.

Improve this answer
$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.