GNU social JP
  • FAQ
  • Login
GNU social JPは日本のGNU socialサーバーです。
Usage/ToS/admin/test/Pleroma FE
  • Public

    • Public
    • Network
    • Groups
    • Featured
    • Popular
    • People

Conversation

Notices

  1. Embed this notice
    lain (lain@fediffusion.art)'s status on Sunday, 29-Oct-2023 20:08:12 JST lain lain
    On the subject of 'nobody really understands how SD works', here's a recent paper that says that every SD model has been trained wrong, which reduces it's capability to produces images that differ from average brightness:

    https://arxiv.org/pdf/2305.08891.pdf

    This is the flaw that something like noise offset tries to correct, but this paper suggest a real fix.
    In conversation Sunday, 29-Oct-2023 20:08:12 JST from fediffusion.art permalink

    Attachments


    • Embed this notice
      Sexy Moon (moon@shitposter.club)'s status on Sunday, 29-Oct-2023 20:08:11 JST Sexy Moon Sexy Moon
      in reply to
      @lain seems like guizzy understands it

      I am trying to decide if I want to learn more in depth how it works or if that will help me in any way toward using it better.
      In conversation Sunday, 29-Oct-2023 20:08:11 JST permalink
    • Embed this notice
      lain (lain@fediffusion.art)'s status on Sunday, 29-Oct-2023 20:13:45 JST lain lain
      in reply to
      • Sexy Moon
      @Moon i can recommend this video series https://www.youtube.com/watch?v=_7rMfsA24Ls
      In conversation Sunday, 29-Oct-2023 20:13:45 JST permalink

      Attachments

      1. Lesson 9: Deep Learning Foundations to Stable Diffusion, 2022
        from Jeremy Howard
        (All lesson resources are available at http://course.fast.ai.) This is the first lesson of part 2 of Practical Deep Learning for Coders. It starts with a tut...
      Sexy Moon likes this.

Feeds

  • Activity Streams
  • RSS 2.0
  • Atom
  • Help
  • About
  • FAQ
  • TOS
  • Privacy
  • Source
  • Version
  • Contact

GNU social JP is a social network, courtesy of GNU social JP管理人. It runs on GNU social, version 2.0.2-dev, available under the GNU Affero General Public License.

Creative Commons Attribution 3.0 All GNU social JP content and data are available under the Creative Commons Attribution 3.0 license.