BUAHAHAHAHAHAHAHAHAH

Are you simulating All Tomorrows?

What is this?

@champion01 three choices what it can be:
A) thing that does iterative calculations with numerical methods/analysis
B) training neural net
C) something else

Pretty sure that is ML training result. but... where is the n-fold validation?

Did someone test using the training set?

I’m not ashamed to say it:
Wat?!

Explain.

@darkwind its B.
I'm tracking specific fixed points per epoch to see how close they start getting with different optimizers and arrangements of training data.

The first delta column is the starting difference between value and inference at the beginning of the training session

the second delta is the epoch's different.

absolute is the distance away from 0, which is what you want it to be.

overall movement is how much the value changed overall that the model returned from the starting epoch to the present one.

Fixed point change indicates how the inferenced result's error changed, going up means it got worse, going down means it got different etc

and absolute movement index is the sum of all absolute change suggestive of how much cumulative difference in error the last epoch made.

@magicMirror that’s my own progress report and no it’s inferencing fixed points not in the training data