Commits · f0a6551292faf8aca40b98c0d2d3b08686b1ad78 · Tamino Huxohl / mu-map · GitLab

Snippets Groups Projects

Jan 04, 2023
- formatting · f0a65512
  Tamino Huxohl authored 2 years ago
  
  f0a65512
- small changes to abstract training: optim.step can be overwritte and logger is truely optional · 68e49293
  Tamino Huxohl authored 2 years ago
  
  68e49293
Dec 22, 2022
- update everything to use key from InterfileKeys class now · e536c7b3
  Tamino Huxohl authored 2 years ago
  
  e536c7b3
Dec 16, 2022
- fix format specifier in snapshot storing · bb734f77
  Tamino Huxohl authored 2 years ago
  
  bb734f77
- fix call to store_snapshot · 42a1436a
  Tamino Huxohl authored 2 years ago
  
  42a1436a
- fix l1 loss usage · 97a7518f
  Tamino Huxohl authored 2 years ago
  
  97a7518f
- fix another bug in eval method · 08480c2a
  Tamino Huxohl authored 2 years ago
  
  08480c2a
- eval_epoch no longer requires the phase argument · 2da7d4f2
  Tamino Huxohl authored 2 years ago
  
  2da7d4f2
- 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 15, 2022
- add missing training lib module · 7bf09ddd
  Tamino Huxohl authored 2 years ago
  
  7bf09ddd
Dec 07, 2022
- update training code · ce73746b
  Tamino Huxohl authored 2 years ago
  
  ce73746b
Nov 10, 2022
- add utilities for interfile and dicom handling · 3b541025
  Tamino Huxohl authored 2 years ago
  
  3b541025
Nov 09, 2022
- add missing __init__.py files · c999a6ce
  Tamino Huxohl authored 2 years ago
  
  c999a6ce
Oct 26, 2022
- add functions to get values from parameter dict to random search · 68689ecc
  Tamino Huxohl authored 2 years ago
  
  68689ecc
- reduce default random search iterations · 107099ef
  Tamino Huxohl authored 2 years ago
  
  107099ef
Oct 25, 2022
- change some random search params again · b6736991
  Tamino Huxohl authored 2 years ago
  
  b6736991
Oct 24, 2022
- test a few different parameters for random search · 29db38c6
  Tamino Huxohl authored 2 years ago
  
  29db38c6
Oct 21, 2022
- add class for random search of best parameters · cdc7ac16
  Tamino Huxohl authored 2 years ago
  
  cdc7ac16
- fix default logger in cgan training and default dataset · 311121fb
  Tamino Huxohl authored 2 years ago
  
  311121fb
- fixed for cgan training · 7981050c
  Tamino Huxohl authored 2 years ago
  
  7981050c
- fix cgan training and patch dataset · 8f7bbd3a
  Tamino Huxohl authored 2 years ago
  
  8f7bbd3a
- update default dataset for cgan training · 56e09b73
  Tamino Huxohl authored 2 years ago
  
  56e09b73
- update conditional gan training for clearer parameters · ed717e73
  Tamino Huxohl authored 2 years ago
  
  ed717e73
Oct 18, 2022
- adapt params of cgan training and measures script uses args correctly · 9b9ef09f
  Tamino Huxohl authored 2 years ago
  
  9b9ef09f
Oct 13, 2022
- use padcrop transform in cgan training · abe29b30
  Tamino Huxohl authored 2 years ago
  
  abe29b30
Oct 12, 2022
- change cgan discriminator back to default · 235fb988
  Tamino Huxohl authored 2 years ago
  
  235fb988
Oct 11, 2022
- add patch discriminator and use it in GAN Training · c2967e1b
  Tamino Huxohl authored 2 years ago
  
  c2967e1b
- fixes to cgan training · 57d641f4
  Tamino Huxohl authored 2 years ago
  
  57d641f4
- make changes in an attempt to fix cgan training · 3536ac24
  Tamino Huxohl authored 2 years ago
  
  3536ac24
- cgan training allows to init generator weights · 5f983873
  Tamino Huxohl authored 2 years ago
  
  5f983873
Oct 10, 2022
- make loss weights in cgan loss configurable · afdee20b
  Tamino Huxohl authored 2 years ago
  
  afdee20b
- make use of the weighted loss function in default training and log floats in more detail · e0e94b14
  Tamino Huxohl authored 2 years ago
  
  e0e94b14
- implement parsing a weighted loss function form a string for CLI usage · de30c17e
  Tamino Huxohl authored 2 years ago
  
  de30c17e
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
- update gan training loss weights · 90df91cb
  Tamino Huxohl authored 2 years ago
  
  90df91cb
- add gdl loss to default training · 249fa333
  Tamino Huxohl authored 2 years ago
  
  249fa333
- update default batch size for cgan training · 450d9f96
  Tamino Huxohl authored 2 years ago
  
  450d9f96
Oct 05, 2022
- implement conditional GAN training · 92f48bfb
  Tamino Huxohl authored 2 years ago
  
  92f48bfb
- fix a bug in args definition for training · 26bc3d50
  Tamino Huxohl authored 2 years ago
  
  26bc3d50