Skip to content
Permalink

Comparing changes

Choose two branches to see what’s changed or to start a new pull request. If you need to, you can also or learn more about diff comparisons.

Open a pull request

Create a new pull request by comparing changes across two branches. If you need to, you can also . Learn more about diff comparisons here.
base repository: CompVis/stable-diffusion
Failed to load repositories. Confirm that selected base ref is valid, then try again.
Loading
base: main
Choose a base ref
...
head repository: Doggettx/stable-diffusion
Failed to load repositories. Confirm that selected head ref is valid, then try again.
Loading
compare: autocast-improvements
Choose a head ref
Checking mergeability… Don’t worry, you can still create the pull request.
  • 13 commits
  • 5 files changed
  • 2 contributors

Commits on Sep 5, 2022

  1. Update attention.py 

    Run attention in a loop to allow for much higher resolutions (over 1920x1920 on a 3090)
    @Doggettx
    Doggettx committed Sep 5, 2022
    Configuration menu
    Copy the full SHA
    afaefd2 View commit details
    Browse the repository at this point in the history

  2. Update attention.py 

    Correction to comment
    @Doggettx
    Doggettx committed Sep 5, 2022
    Configuration menu
    Copy the full SHA
    5065b41 View commit details
    Browse the repository at this point in the history

  3. Update attention.py

    @Doggettx
    Doggettx committed Sep 5, 2022
    Configuration menu
    Copy the full SHA
    8283bb5 View commit details
    Browse the repository at this point in the history

  4. Fixed memory handling for model.decode_first_stage 

    Better memory handling for model.decode_first_stage so it doesn't crash anymore after 100% rendering
    @Doggettx
    Doggettx committed Sep 5, 2022
    Configuration menu
    Copy the full SHA
    d3c91ec View commit details
    Browse the repository at this point in the history

Commits on Sep 6, 2022

  1. Fixed free memory calculation 

    Old version gave incorrect free memory results causing in crashes on edge cases.
    @Doggettx
    Doggettx committed Sep 6, 2022
    Configuration menu
    Copy the full SHA
    507ddec View commit details
    Browse the repository at this point in the history

  2. Set model to half 

    Set model to half in txt2img and img2img for less memory usage.
    @Doggettx
    Doggettx committed Sep 6, 2022
    Configuration menu
    Copy the full SHA
    a1fbe55 View commit details
    Browse the repository at this point in the history

  3. Commented out debug info 

    Forgot to comment out debug info
    @Doggettx
    Doggettx committed Sep 6, 2022
    Configuration menu
    Copy the full SHA
    7a32fd6 View commit details
    Browse the repository at this point in the history

  4. Raise error when steps too high 

    Technically you could run at higher steps as long as the resolution is dividable by the steps but you're going to run into memory issues later on anyhow.
    @Doggettx
    Doggettx committed Sep 6, 2022
    Configuration menu
    Copy the full SHA
    755bec8 View commit details
    Browse the repository at this point in the history

  5. Added max. res info to memory exception

    @Doggettx
    Doggettx committed Sep 6, 2022
    Configuration menu
    Copy the full SHA
    f134e24 View commit details
    Browse the repository at this point in the history

Commits on Sep 7, 2022

  1. Reverted in place tensor functions back to CompVis version 

    Improves performance and is no longer needed.
    @Doggettx
    Doggettx committed Sep 7, 2022
    Configuration menu
    Copy the full SHA
    830f694 View commit details
    Browse the repository at this point in the history

  2. Missed one function to revert

    @Doggettx
    Doggettx committed Sep 7, 2022
    2 Configuration menu
    Copy the full SHA
    c2d72c5 View commit details
    Browse the repository at this point in the history

  3. Update README.md

    @Doggettx
    Doggettx authored Sep 7, 2022
    Configuration menu
    Copy the full SHA
    cd3d653 View commit details
    Browse the repository at this point in the history

Commits on Sep 10, 2022

  1. Performance boost and fix sigmoid for higher resolutions 

    Significant performance boost at higher resolutions when running in auto_cast or half mode

on 3090 went from 1.13it/s to 1.63it/s at 1024x1024

Will also allow for higher resolutions due to sigmoid fix and using half memory
    @Doggettx
    Doggettx committed Sep 10, 2022
    Configuration menu
    Copy the full SHA
    5fe97c6 View commit details
    Browse the repository at this point in the history
Loading