Commit Graph

6 Commits

Author SHA1 Message Date
xzuyn
91ab594744 minor fix, some changes, some additions, and cleaning up (#1618)
* - fix overflowing text (a janky fix)
- add DEISMultistep scheduler as an option
- set default scheduler to DEISMultistep
- set default CFG to 3.5
- set default steps to 16
- add `xzuyn/PhotoMerge` as a model option
- add 3 new example prompts (which work nicely with PhotoMerge)
- formatting

* Set DEISMultistep in the cpu_only list instead

* formatting

* formatting

* modify prompts

* resize window to 81% & 85% monitor resolution instead of (WxH / 1.0625).

* increase steps to 32 after some testing. somewhere in between 16 and 32 is best compromise on speed/quality for DEIS, so 32 steps to play it safe.

* black line length 79

* revert settings DEIS as default scheduler.

* add more schedulers & revert accidental DDIM change
- add DPMSolverSingleStep, KDPM2AncestralDiscrete, & HeunDiscrete.
- did not add `DPMSolverMultistepInverse` or `DDIMInverse` as they only output latent noise, there are a few I did not try adding yet.
- accidentally set `upscaler_ui.py` to EulerDiscrete by default last commit while reverting DEIS changes.
- add `xzuyn/PhotoMerge-inpainting` as an in or out painting model.

* black line length 79

* add help section stuff and some other changes
- list the rest of the schedulers in argument help section.
- replace mutable default arguments.
- increased default window height to 91% to remove any scrolling for the main txt2img page (tested on a 1920x1080 monitor). width is the same as its just enough to have the image output on the side instead of the bottom.
- cleanup
2023-07-04 18:51:23 -07:00
Kyle Herndon
606029c01c Fix LoRA device format bug and allow LoRA to resume from a previous training 2023-04-17 13:19:46 +05:30
Kyle Herndon
0b0526699a Fix incorrect device argument initialization for LoRA training by extracting the device type and number and formatting it for pytorch (#1237)
Co-authored-by: Kyle Herndon <kyle@nod-labs.com>
2023-03-24 01:10:50 -07:00
jinchen62
420366c1b8 Move schedulers to global obj (#1225) 2023-03-21 22:40:43 -07:00
Kyle Herndon
c23fcf3748 Fix incorrect device argument initialization for LoRA training (#1231)
Co-authored-by: Kyle Herndon <kyle@nod-labs.com>
Co-authored-by: powderluv <powderluv@users.noreply.github.com>
2023-03-21 19:07:18 -07:00
Kyle Herndon
ba48ff8d25 Implement LoRA training and UI for training and UI for inference in img2img, inpaint, outpaint (#1200)
txt2img inference UI is already committed.

Co-authored-by: Kyle Herndon <kyle@nod-labs.com>
Co-authored-by: powderluv <powderluv@users.noreply.github.com>
2023-03-17 12:54:56 -07:00