Commits · 8b7500042547b1fb71c1eb10c464a09cfdd57ea7 · Tamino Huxohl / mu-map · GitLab

Snippets Groups Projects

Oct 06, 2022
- update loss curve vis by adding new parameters · 8b750004
  Tamino Huxohl authored 2 years ago
  
  8b750004
- 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
- update gitignore · 148c1d80
  Tamino Huxohl authored 2 years ago
  
  148c1d80
Oct 05, 2022
- update discriminator to use a sigmnoid in the end · 06a38be4
  Tamino Huxohl authored 2 years ago
  
  06a38be4
- 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
- implement new patch dataset parameters into training · 6715925a
  Tamino Huxohl authored 2 years ago
  
  6715925a
- fix parameter updates to patch dataset · 82eb0bf4
  Tamino Huxohl authored 2 years ago
  
  82eb0bf4
- update parameters to patches dataset · c4ac5562
  Tamino Huxohl authored 2 years ago
  
  c4ac5562
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
- update gitignore · f859b9d0
  Tamino Huxohl authored 2 years ago
  
  f859b9d0
- fix that transforms are correctly passed to the dataset · 338520da
  Tamino Huxohl authored 2 years ago
  
  338520da
- patch dataset now correctly uses augemntation transforms · b8341337
  Tamino Huxohl authored 2 years ago
  
  b8341337
- 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
- implement a scale tranform · e95e5b6f
  Tamino Huxohl authored 2 years ago
  
  e95e5b6f
- default training now logs args · 58c38ae5
  Tamino Huxohl authored 2 years ago
  
  58c38ae5
- update the loss curve script · 686d2316
  Tamino Huxohl authored 2 years ago
  
  686d2316
- update parameters in default training · a9a61785
  Tamino Huxohl authored 2 years ago
  
  a9a61785
- test batch size for patches in unet main · 2c050de5
  Tamino Huxohl authored 2 years ago
  
  2c050de5
Sep 28, 2022
- introduce parameters to patch dataset contructor · fcb23a3e
  Tamino Huxohl authored 2 years ago
  
  fcb23a3e
- Merge branch 'main' of https://gitlab.ub.uni-bielefeld.de/thuxohl/mu-map · 91b8bd38
  Tamino Huxohl authored 2 years ago
  
  91b8bd38
- update parameters and dataset in default training · d961d0a5
  Tamino Huxohl authored 2 years ago
  
  d961d0a5
- Merge branch 'main' of https://gitlab.ub.uni-bielefeld.de/thuxohl/mu-map · db4917fa
  Tamino Huxohl authored 2 years ago
  
  db4917fa
Sep 27, 2022
- mu map dataset can now load splits · ac263a4e
  Tamino Huxohl authored 2 years ago
  
  ac263a4e
- splits are now parsed as dicts · 6442ed68
  Tamino Huxohl authored 2 years ago
  
  6442ed68
- make default training configurable via argparse · b36a8455
  Tamino Huxohl authored 2 years ago
  
  b36a8455
- implement visualization of results in test script · d66fd526
  Tamino Huxohl authored 2 years ago
  
  d66fd526
- mock dataset normalizes to maximum mu_map coefficient in the dataset · 9beb28b1
  Tamino Huxohl authored 2 years ago
  
  9beb28b1
- max normalization allows to define norm values · 63680f4b
  Tamino Huxohl authored 2 years ago
  
  63680f4b
- add capability to parse logfiles and generate loss curves from that · e44371f2
  Tamino Huxohl authored 2 years ago
  
  e44371f2
- update mock dataset module by re-using display/main function from default dataset · 9d4011f3
  Tamino Huxohl authored 2 years ago
  
  9d4011f3
- add transforms to default dataset · bed5ba58
  Tamino Huxohl authored 2 years ago
  
  bed5ba58
- fix for normalization · b43e8bce
  Tamino Huxohl authored 2 years ago
  
  b43e8bce
- add some documentation for transforms · ee052e16
  Tamino Huxohl authored 2 years ago
  
  ee052e16
- formatting · 85930b5e
  Tamino Huxohl authored 2 years ago
  
  85930b5e
- add abstract transform module · 552a1ae3
  Tamino Huxohl authored 2 years ago
  
  552a1ae3
- restructuring to separate data processing from torch dataset implementations · 2ae5e9e3
  Tamino Huxohl authored 2 years ago
  
  2ae5e9e3