Timeline for Training loss goes down and up again. What is happening?
Current License: CC BY-SA 3.0
8 events
| when toggle format | what | by | license | comment | |
|---|---|---|---|---|---|
| Apr 8, 2016 at 13:21 | vote | accept | patapouf_ai | ||
| Apr 8, 2016 at 12:20 | answer | added | itdxer | timeline score: 31 | |
| Apr 5, 2016 at 15:43 | comment | added | patapouf_ai | Thank you itdxer. I think what you said must be on the right track. I tried using "adam" instead of "adadelta" and this solved the problem, though I'm guessing that reducing the learning rate of "adadelta" would probably have worked also. If you want to write a full answer I shall accept it. | |
| Mar 11, 2016 at 15:51 | comment | added | itdxer | Sorry, I mean learning rate. | |
| Mar 11, 2016 at 15:49 | comment | added | patapouf_ai | But how could extra training make the training data loss bigger? | |
| Mar 11, 2016 at 13:17 | comment | added | itdxer | Your learning could be to big after the 25th epoch. Try to set up it smaller and check your loss again | |
| Mar 11, 2016 at 10:23 | review | First posts | |||
| Mar 11, 2016 at 10:47 | |||||
| Mar 11, 2016 at 10:18 | history | asked | patapouf_ai | CC BY-SA 3.0 |