dodox/shark
1
0
Fork 0
Code Issues Pull Requests Releases Wiki Activity

Commit Graph

  • cd89945d55 bs-constant now only in one place and updated README master dodox 2021-09-22 22:27:53 +02:00
  • 4f250f61c3 Smol changes on parameters dodox 2021-09-22 18:53:56 +02:00
  • d7c5d57d93 Decreased blob-size a bit to make training faster dodox 2021-09-22 12:10:17 +02:00
  • ca964732c5 Removed the slow blob-len growth dodox 2021-09-22 11:50:03 +02:00
  • 8608a89909 Train on bigger blobs (1KiB) dodox 2021-09-22 11:46:22 +02:00
  • e4e2a18606 Tweaked training params dodox 2021-09-22 11:09:15 +02:00
  • c95d5a2e8b Merge branch 'master' of dominik.roth.ml:dodox/shark dodox 2021-09-22 10:38:39 +02:00
  • 483201c693 Increased layer count of model and refactored model into its own file dodox 2021-09-22 10:37:11 +02:00
  • 18ba954907 Increased layer count of model dodox 2021-09-22 10:37:11 +02:00
  • 1b4b6f1a2f Updated README and better epoch-printing dodox 2021-09-22 10:28:08 +02:00
  • 5f0f9f1dce Fixed a problem in model-training and renamed files dodox 2021-09-22 09:14:23 +02:00
  • 30cc846c6f Sucessful discrimination; updated README dodox 2021-09-21 16:00:43 +02:00
  • 42ed2dd676 Fix: Encryption did not correctly increment iv dodox 2021-09-21 15:54:29 +02:00
  • 0c197e3e56 Higher lr; still wont work (gives false in 100% of tests) dodox 2021-09-21 12:57:55 +02:00
  • 054b4494d2 Even better loss tracking and more sparse checkpointing dodox 2021-09-21 11:05:28 +02:00
  • 5bd50e2b43 Better loss tracking and implemented saving of network-wheights dodox 2021-09-21 09:49:27 +02:00
  • 627bf370bc Upped lr and dropout dodox 2021-09-21 09:17:01 +02:00
  • f4444f3a9e Started implementation of the PPT-Attacker dodox 2021-09-21 09:14:31 +02:00
  • 5f122e3cbf The way markdown handles newlines is stupidwq dodox 2021-09-20 11:35:29 +02:00
  • 1eeeee57fe small README dodox 2021-09-20 11:34:57 +02:00
  • 0dabde6dad initial commit dodox 2021-09-20 11:32:55 +02:00