So I've started looking into LSTMs, and I had a doubt. Let's say I have data for 100 days, with each day's data being 1000x1000. So the input data is a 3D matrix of dimension 1000x1000x100. I want to do batch learning for every day's data. I have 10 batches of size 100. Now my question is: wouldn't doing batch learning like this affect the sequence of the data? I'm doing 10 batches for each day by inputting one after another, essentially making it look like 10 days of data, right? So now it does not look like one day's data. Any help regarding this would be greatly appreciated :)
This is for a project I'm doing, and the input data is way larger than this. I was wondering on how to split data for each day.