Timeline for When to stop training of neural network when validation loss is still decreasing but gap with training loss is increasing?
Current License: CC BY-SA 4.0
6 events
when toggle format | what | by | license | comment | |
---|---|---|---|---|---|
Sep 5, 2019 at 3:11 | answer | added | kbrose | timeline score: 1 | |
Oct 31, 2018 at 6:00 | history | tweeted | twitter.com/StackStats/status/1057512613326807040 | ||
Jul 13, 2018 at 15:01 | comment | added | Matthew Drury | Sure, will do after i get off the bus! | |
Jul 13, 2018 at 14:37 | comment | added | Soltius | Could you please elaborate @MatthewDrury ? I know a small gap is nothing to worry about, but what does a widening gap like so indicate ? | |
Jul 13, 2018 at 14:34 | comment | added | Matthew Drury | A gap between training and testing loss does not indicate overfitting. This is a very common fallacy. | |
Jul 13, 2018 at 14:29 | history | asked | Soltius | CC BY-SA 4.0 |