Considering batch size=1 implies online machine learning??
What is the difference between an online machine learning model and traditional machine learning model with batch size=1
$\begingroup$
$\endgroup$
Add a comment
|
1 Answer
$\begingroup$
$\endgroup$
There are many ways to train models in online mode, but in general this would mean that you update your model as the data arrives. In non-online mode, you iterate over all the data, while in online mode this could not be possible. In non-online mode, even with batch size of one sample and using something like stochastic gradient descent, you would usually train the model for many epochs, i.e. loop over the batches many times. In online mode you usually don’t go back to past samples.
