Problem in training about anfistensorflow2.0

[gregorLen]

Hi Mustafa! Thanks alot for the feedback! I will look into it!

[Mustafa-ismail-Eyi]

Hello, i was working with this repo for my task for a few days. I think, I figure out the problem. Problem is caused from 2 thing:

1-) You have to scale your features and your target into 0-1. Be sure you do that
2-) I don't know why but the batch size causes it. Decrease the batch size from 16 to 8 or less

If you don't do step 1, your learning progress something like:
loss: nan, val_loss: nan

[Metri-Co]

I was working this repo and you must match the batch number in 'param' and the batch size in split_data function

I ran the code with batch size of 16 and i had no problem. Hope it works.

[feyzazeynep]

Hello, we can estimate the data we have. So how do we predict for the future?

[gregorLen]

see answer in #5 (comment)