* - 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
* Adding metal_utils for iree_utils
* Add patch for making compile API work for both MEGABYTE and MiniGPT4 (#1559)
-- It also modifies the mega_test.py script
Signed-off-by: Abhishek Varma <abhishek@nod-labs.com>
* [SD] Update unet in_channels API and add PIL metadata to spec. (#1560)
* Fix deprecation warning for unet config.
* Include PIL metadata instead of hidden imports in SD spec.
* Fixing iree-metal-target-platform
* adding metal to txt2img pipeline
* Fixing Copyright date
* removing debug prints
* black lint formating
* fixing device dump
---------
Signed-off-by: Abhishek Varma <abhishek@nod-labs.com>
Co-authored-by: Abhishek Varma <avarma094@gmail.com>
Co-authored-by: Ean Garvey <87458719+monorimet@users.noreply.github.com>
Co-authored-by: powderluv <powderluv@users.noreply.github.com>
* Change retrieval of Iree options used when loading an existing
unet512 vmfb to look up the "unet" options rather than attempt to
find a non-existent set of options for "unet512"
Co-authored-by: powderluv <powderluv@users.noreply.github.com>
* Switches a few places in the SD pipeline where an assumption of
max_length=64 was being made, to using the actual max_length
as passed into the pipeline. This prevents unet512 always being
used and producing different images than previously when
--max_length=77
* Fixes randomize seed buttons that stopped working.
* Update now deprecated method to set initial colums for output
gallery to the newer undeprecated one.
* On startup report that cleaning up of temp files is taking place, in
case it takes a long time.
* Have the output gallery tab delete any zero length temporary files
generated by gradio < 3.32.0 for its gallery control whenever it
needs to update that control with images. This prevents such
files multiplying out of control.
This fix the gallery component of the output gallery dumping temporary
files into the standard folders rather than shark_tmp so those files never
got cleared out on restart and would build up.
