Commits · 5d9d695e6de9688bb518cbb6f3cefffa791aa351 · Tamino Huxohl / mu-map · GitLab

Snippets Groups Projects

Jan 13, 2023
- update distance training to use early stopping · 5d9d695e
  Tamino Huxohl authored 2 years ago
  
  5d9d695e
Jan 04, 2023
- documen training implementations · e3e6d7d5
  Tamino Huxohl authored 2 years ago
  
  e3e6d7d5
- renaming of distance training class · 9a9d4100
  Tamino Huxohl authored 2 years ago
  
  9a9d4100
Dec 22, 2022
- update everything to use key from InterfileKeys class now · e536c7b3
  Tamino Huxohl authored 2 years ago
  
  e536c7b3
Dec 16, 2022
- fix l1 loss usage · 97a7518f
  Tamino Huxohl authored 2 years ago
  
  97a7518f
- distance training moves model to correct device · 2f7abaaa
  Tamino Huxohl authored 2 years ago
  
  2f7abaaa
- rename default training into distance training and fix issues with usage as abstract training · 837a67db
  Tamino Huxohl authored 2 years ago
  
  837a67db
Dec 07, 2022
- update training code · ce73746b
  Tamino Huxohl authored 2 years ago
  
  ce73746b
Oct 10, 2022
- make use of the weighted loss function in default training and log floats in more detail · e0e94b14
  Tamino Huxohl authored 2 years ago
  
  e0e94b14
Oct 07, 2022
- remove output scale param from trainings as this is now done by rescaling the precision factor · ce61a834
  Tamino Huxohl authored 2 years ago
  
  ce61a834
Oct 06, 2022
- add gdl loss to default training · 249fa333
  Tamino Huxohl authored 2 years ago
  
  249fa333
Oct 05, 2022
- fix a bug in args definition for training · 26bc3d50
  Tamino Huxohl authored 2 years ago
  
  26bc3d50
- implement new patch dataset parameters into training · 6715925a
  Tamino Huxohl authored 2 years ago
  
  6715925a
Oct 04, 2022
- change loss from MSE to L1 · e8f15d80
  Tamino Huxohl authored 2 years ago
  
  e8f15d80
- remove a debug comment from training · 5cf43e4d
  Tamino Huxohl authored 2 years ago
  
  5cf43e4d
- fix that transforms are correctly passed to the dataset · 338520da
  Tamino Huxohl authored 2 years ago
  
  338520da
- implement initialization of random seed · 455bed65
  Tamino Huxohl authored 2 years ago
  
  455bed65
- implement selection of normalization and scale in default training · 102f3e46
  Tamino Huxohl authored 2 years ago
  
  102f3e46
- default training now logs args · 58c38ae5
  Tamino Huxohl authored 2 years ago
  
  58c38ae5
- update parameters in default training · a9a61785
  Tamino Huxohl authored 2 years ago
  
  a9a61785
Sep 28, 2022
- update parameters and dataset in default training · d961d0a5
  Tamino Huxohl authored 2 years ago
  
  d961d0a5
Sep 27, 2022
- make default training configurable via argparse · b36a8455
  Tamino Huxohl authored 2 years ago
  
  b36a8455
Sep 26, 2022
- fix usage of snapshot epoch param in training · 96e9d37b
  Tamino Huxohl authored 2 years ago
  
  96e9d37b
- first implementation of default training · aadd524d
  Tamino Huxohl authored 2 years ago
  
  aadd524d
Sep 01, 2022
- add a code for training · 0cc09b15
  Tamino Huxohl authored 2 years ago
  
  0cc09b15