Allows setting of preferred VAE and Text encoder(s) for checkpoints when selected via Checkpoint cards. No selection saved means no change to current toprow setting. 'Built in' option, if the only choice, means clear the toprow selection (therefore use vae/te built-in to checkpoint).
Also allows setting model type for checkpoints (SD1/SD2/SDXL/Flux/Unknown) (user set only, no attempt at autodetection), enabling filtering of the cards based on UI preset.
adds options for user-set defaults for Sampler and Scheduler to UI settings sd, xl, flux;
adds options for user-set defaults for GPU Weights to UI settings xl, flux;
necessitates change to .input event listener instead of .release for ui_forge_inference_memory, which may be more correct anyway.
Added Flux to lora types in extra networks UI, so user can set.
Loras versioned first by user-set type, if any. Falls back to heuristics - these are much more reliable than the removed old A1111 tests and in case of no match default to Unknown (always displayed).
Filtering is done based on UI setting. 'all' setting does not filter. Filters lora lists on change.
Removed unused 'lora_hide_unknown_for_versions' setting.
Previously these were Numbers, and limited to integers, with no range check.
Some schedulers don't respect these values anyway, but that's a different issue.
* autoset width and height
When loading into img2img, added an option to autoset Width and Height from image (Settings -> img2img ->After loading into Img2img, automatically update Width and Height).
Also fixed scale_by display, now correctly rounded to multiple of 8
Also fixed img2img new width/height calculation, same way, (affects infotext)
* Update ui.py: hide progress animation on image size change
Now doesn't spam clicks, so cancel will stop.
So works correctly when swapping tabs within webui.
But still won't continue generating if switch browser tab or minimise browser. It's like the click doesn't get processed. Possibly related to notification sound not playing if tab not active.
#1760, original solution by cmdr2. extended to 2 other locations where setting is read. I could replicate the issue only by manually entering a decimal value.
Possibly also #1764. Though I don't know how the clipskip setting became a string.
* SAG updates
* restore settings from infotext
* bypass if model is Flux (would error during generation)
* bypass if CFG == 1 (would error during generation)
* add control of blur mask threshold (previously hardcoded)
* change to InputAccordion
* forge_sag.py: update title
Re-enables *hires_fix_show_sampler* and *hires_fix_show_prompts* user-configurable settings.
Also corrects the reported result width x height; now rounded down to multiple of 8.
#1247
* Update aspectRatioOverlay.js
fix finding the right element to draw the box on
add drawing box based on changes to number control, just not slider
* Update style.css
add margin equal to border, to draw the AR box in better position
* fix GFPGAN to work with visibility < 1
* fix codeformer to work with visibility < 1
* try harder to download GFPGAN model. Old method would download only if there were no .pth models in the GFPGAN directory. If codeformer was used before GFPGAN, the supporting models are already downloaded into the GFPGAN directory.
The previous change to the width of the input box, to prevent overlap between long checkpoint names and the dropdown arrow, also reduced the clickable region of the input box - showing up the arrow as the purely visual element that it is.
This PR restores width, tweaks position of the arrow, and makes the arrow background solid to avoid inference with text. The latter two changes also affect the vae/te and diffusion in low bits menus.
* Fix Checkpoint Merging #1359,#1095
- checkpoint_list[] contains the CheckpointInfo.title which is "checkpointname.safetensor [hash]"
when a checkpoint is selected to be loaded during merge, we try to match it with just "checkpointname.safetensor".
-> use checkpoint_aliases[] which already contains the checkpoint key in all possible variants.
- replaced removed sd_models.read_state_dict() with sd_models.load_torch_file()
- replaced removed sd_vae.load_vae_dict() with sd_vae.load_torch_file()
- uncommented create_config() for now, since it calls a removed method: sd_models_config.find_checkpoint_config_near_filename()
* Follow up merge fix for #1359#1095
- read_state_dict() does nothing, replaced 2 occurrences with load_torch_file()
- now merging actually merges again
