1.6.0-RC

[AUTOMATIC1111]

To use this pre-release, you must use the release_candidate branch.

To get the release_candidate branch in a new webui installation, run those commands (this will create a directory called stable-diffusion-webui):

git clone https://github.com/AUTOMATIC1111/stable-diffusion-webui.git
cd stable-diffusion-webui
git switch release_candidate
webui-user.bat

Alternatively, to switch your existing installation to release_candidate branch:

git switch release_candidate
git pull

To switch back to master branch:

git switch master

Features:

• refiner support #12371
• add NV option for Random number generator source setting, which allows to generate same pictures on CPU/AMD/Mac as on NVidia videocards
• add style editor dialog
• hires fix: add an option to use a different checkpoint for second pass (#12181)
• option to keep multiple loaded models in memory (#12227)
• new samplers: Restart, DPM++ 2M SDE Exponential, DPM++ 2M SDE Heun, DPM++ 2M SDE Heun Karras, DPM++ 2M SDE Heun Exponential, DPM++ 3M SDE, DPM++ 3M SDE Karras, DPM++ 3M SDE Exponential (#12300, #12519, #12542)
• rework DDIM, PLMS, UniPC to use CFG denoiser same as in k-diffusion samplers:
  • makes all of them work with img2img
  • makes prompt composition posssible (AND)
  • makes them available for SDXL
• always show extra networks tabs in the UI (#11808)
• use less RAM when creating models (#11958, #12599)
• textual inversion inference support for SDXL
• extra networks UI: show metadata for SD checkpoints
• checkpoint merger: add metadata support
• prompt editing and attention: add support for whitespace after the number ([ red : green : 0.5 ]) (seed breaking change) (#12177)
• VAE: allow selecting own VAE for each checkpoint (in user metadata editor)
• VAE: add selected VAE to infotext
• options in main UI: add own separate setting for txt2img and img2img, correctly read values from pasted infotext, add setting for column count (#12551)
• add resize handle to txt2img and img2img tabs, allowing to change the amount of horizontable space given to generation parameters and resulting image gallery (#12687, #12723)
• change default behavior for batching cond/uncond -- now it's on by default, and is disabled by an UI setting (Optimizatios -> Batch cond/uncond) - if you are on lowvram/medvram and are getting OOM exceptions, you will need to enable it
• show current position in queue and make it so that requests are processed in the order of arrival (#12707)
• add --medvram-sdxl flag that only enables --medvram for SDXL models
• prompt editing timeline has separate range for first pass and hires-fix pass (seed breaking change) (#12457)

Minor:

• img2img batch: RAM savings, VRAM savings, .tif, .tiff in img2img batch (#12120, #12514, #12515)
• postprocessing/extras: RAM savings (#12479)
• XYZ: in the axis labels, remove pathnames from model filenames
• XYZ: support hires sampler (#12298)
• XYZ: new option: use text inputs instead of dropdowns (#12491)
• add gradio version warning
• sort list of VAE checkpoints (#12297)
• use transparent white for mask in inpainting, along with an option to select the color (#12326)
• move some settings to their own section: img2img, VAE
• add checkbox to show/hide dirs for extra networks
• Add TAESD(or more) options for all the VAE encode/decode operation (#12311)
• gradio theme cache, new gradio themes, along with explanation that the user can input his own values (#12346, #12355)
• sampler fixes/tweaks: s_tmax, s_churn, s_noise, s_tmax (#12354, #12356, #12357, #12358, #12375, #12521)
• update README.md with correct instructions for Linux installation (#12352)
• option to not save incomplete images, on by default (#12338)
• enable cond cache by default
• git autofix for repos that are corrupted (#12230)
• allow to open images in new browser tab by middle mouse button (#12379)
• automatically open webui in browser when running "locally" (#12254)
• put commonly used samplers on top, make DPM++ 2M Karras the default choice
• zoom and pan: option to auto-expand a wide image, improved integration (#12413, #12727)
• option to cache Lora networks in memory
• rework hires fix UI to use accordion
• face restoration and tiling moved to settings - use "Options in main UI" setting if you want them back
• change quicksettings items to have variable width
• Lora: add Norm module, add support for bias (#12503)
• Lora: output warnings in UI rather than fail for unfitting loras; switch to logging for error output in console
• support search and display of hashes for all extra network items (#12510)
• add extra noise param for img2img operations (#12564)
• support for Lora with bias (#12584)
• make interrupt quicker (#12634)
• configurable gallery height (#12648)
• make results column sticky (#12645)
• more hash filename patterns (#12639)
• make image viewer actually fit the whole page (#12635)
• make progress bar work independently from live preview display which results in it being updated a lot more often
• forbid Full live preview method for medvram and add a setting to undo the forbidding
• make it possible to localize tooltips and placeholders

Extensions and API:

• gradio 3.41.0
• also bump versions for packages: transformers, GitPython, accelerate, scikit-image, timm, tomesd
• support tooltip kwarg for gradio elements: gr.Textbox(label='hello', tooltip='world')
• properly clear the total console progressbar when using txt2img and img2img from API
• add cmd_arg --disable-extra-extensions and --disable-all-extensions (#12294)
• shared.py and webui.py split into many files
• add --loglevel commandline argument for logging
• add a custom UI element that combines accordion and checkbox
• avoid importing gradio in tests because it spams warnings
• put infotext label for setting into OptionInfo definition rather than in a separate list
• make StableDiffusionProcessingImg2Img.mask_blur a property, make more inline with PIL GaussianBlur (#12470)
• option to make scripts UI without gr.Group
• add a way for scripts to register a callback for before/after just a single component's creation
• use dataclass for StableDiffusionProcessing
• store patches for Lora in a specialized module instead of inside torch
• support http/https URLs in API (#12663, #12698)
• add extra noise callback (#12616)
• dump current stack traces when exiting with SIGINT
• add type annotations for extra fields of shared.sd_model

Bug Fixes:

• Don't crash if out of local storage quota for javascriot localStorage
• XYZ plot do not fail if an exception occurs
• fix missing TI hash in infotext if generation uses both negative and positive TI (#12269)
• localization fixes (#12307)
• fix sdxl model invalid configuration after the hijack
• correctly toggle extras checkbox for infotext paste (#12304)
• open raw sysinfo link in new page (#12318)
• prompt parser: Account for empty field in alternating words syntax (#12319)
• add tab and carriage return to invalid filename chars (#12327)
• fix api only Lora not working (#12387)
• fix options in main UI misbehaving when there's just one element
• make it possible to use a sampler from infotext even if it's hidden in the dropdown
• fix styles missing from the prompt in infotext when making a grid of batch of multiplie images
• prevent bogus progress output in console when calculating hires fix dimensions
• fix --use-textbox-seed
• fix broken Lora/Networks: use old method option (#12466)
• properly return None for VAE hash when using --no-hashing (#12463)
• MPS/macOS fixes and optimizations (#12526)
• add second_order to samplers that mistakenly didn't have it
• when refreshing cards in extra networks UI, do not discard user's custom resolution
• fix processing error that happens if batch_size is not a multiple of how many prompts/negative prompts there are (#12509)
• fix inpaint upload for alpha masks (#12588)
• fix exception when image sizes are not integers (#12586)
• fix incorrect TAESD Latent scale (#12596)
• auto add data-dir to gradio-allowed-path (#12603)
• fix exception if extensuions dir is missing (#12607)
• fix issues with api model-refresh and vae-refresh (#12638)
• fix img2img background color for transparent images option not being used (#12633)
• attempt to resolve NaN issue with unstable VAEs in fp32 mk2 (#12630)
• implement missing undo hijack for SDXL
• fix xyz swap axes (#12684)
• fix errors in backup/restore tab if any of config files are broken (#12689)
• fix SD VAE switch error after model reuse (#12685)
• fix trying to create images too large for the chosen format (#12667)
• create Gradio temp directory if necessary (#12717)
• prevent possible cache loss if exiting as it's being written by using an atomic operation to replace the cache with the new version
• set devices.dtype_unet correctly
• run RealESRGAN on GPU for non-CUDA devices (#12737)
• prevent extra network buttons being obscured by description for very small card sizes (#12745)

Full Changelog: v1.5.1...v1.6.0-RC

---

This discussion was created from the release 1.6.0-RC.

[rltgjqmcpgjadyd]

finally!

[lllyasviel]

The extension sd-webui-controlnet is currently working on supporting some SDXL control models.

Mikubill/sd-webui-controlnet#1952

[szclcwgdx]

"Do you mean that if we take the official control-lora model released by SD and place it in sd-webui-controlnet, it can run?"

[Lling00]

期待支持SDXL 的sd-webui-controlnet

[eniora]

Eagerly waiting for this! Thank you. Tried CN with comfy and it worked fine but it's just too complicated for me 😂

[suonnon]

After using "git switch release_candidate" and "git pull" I get this error:

There is no tracking information for the current branch.
Please specify which branch you want to merge with.
See git-pull(1) for details.

git pull <remote> <branch>

If you wish to set tracking information for this branch you can do so with:

git branch --set-upstream-to=origin/<branch> release_candidate

[AUTOMATIC1111]

try git branch --set-upstream-to=origin/release_candidate release_candidate

[suonnon]

@AUTOMATIC1111 Thank you!!!! 🙏🙏

[hungtooc]

I see there is a NSFW safety checker feature #367 (comment)
but the description doesn't note about it. And how to use them on webui and API too.

[AUTOMATIC1111]

There is an old extension that does censoring: https://github.com/AUTOMATIC1111/stable-diffusion-webui-nsfw-censor

I tested it just now, and it seems to work.

[suonnon]

Every dropdown menus in main UI has a bug:

• It's automatically collapsed/close the dropdown menu when click (right after it was opened)
  

[catboxanon]

Gradio broke this. It'll be fixed in the next version. gradio-app/gradio#5323

Edit: This fix made it's way into Gradio 3.41.1, and 1.6.0 now uses 3.41.2, so this is resolved.

[Kadah]

Dropdowns are super buggy currently. Scrolling on long ones (either the list itself or the page) is pretty unreliable.

Its gotten better in the last couple Gradio update. Its no longer opening an invisible dropdown on its own after selection which will cause an unwanted second selection change. That was super annoying on the checkpoint selection one: change model, go to click in prompt field, model changes again cause there was a ghost checkpoint selection dropdown open.

[suonnon]

Bug => Stuck at "Reloading".

• Whenever I press "Reload UI" from footer button or "Apply Settings and Reload UI"

[skittlz444]

That's been the case for a few versions for me, I think another fork addresses that. Was going to investigate tomorrow.

[ColorfanXP]

What ever happened to civitai helper 2.0? Did they just give up development or something?

[suonnon]

What ever happened to civitai helper 2.0? Did they just give up development or something?

Let's hope they are not 🙏 It's damn must have useful extension!!!

[rltgjqmcpgjadyd]

oops developer removed the icon :(

[ColorfanXP]

Yeah, but hes been inactive for a long while. And civitai helper 2 was only a concept.

[szclcwgdx]

The "Hires. fix" and "Refiner" sections don't have toggle switches, so I'm not sure if they're turned on or off, haha.

[elistys]

@skittlz444
@catboxanon
The ideal and perfect accordion and toggle combination can be found in DaVinci Resolve's interface.

1.mov

[skittlz444]

@elistys yess that would be ideal. Doubt we'll see that anytime soon though

[TrickleDownClown]

Where does style.css reside?

[skittlz444]

Style.css is in the root directory of the project, but it shouldn't be edited directly unless making a PR to the repo, you should create a user.css in the same directory for custom styling

[vxarg]

This also breaks some extensions such as the Queue Extension (https://github.com/Kryptortio/SDAtom-WebUi-client-queue-ext).

Also, the Defaults page under settings doesn't seem to remember the on/off state of Hi-Res fix.
I open and close it and click view changes and and it says "no changes".

So I can only tweak the default settings for Hi-Res Fix (steps, denoising, etc) but can't have Hi-Res Fix turned on by default.

Not sure if I should submit a new bug report for this?

[andzejsp]

Who needs hires fix when you got refiner. Damn it works so good, mo more manual refining, just set a value at which it switches to refiner model and you're golden.
Thank you!

[suonnon]

It's not about the refiner, it's about the UI/UX design.

[andzejsp]

How do you save styles/prompts now? there is no button to save it

[catboxanon]

The paintbrush icon opens an editor for styles -- you can create and save them there.

[suonnon]

Is there a way to turn-off automatic opening the default browser when backend fully load?

[catboxanon]

Settings -> System -> Automatically open webui in browser on startup

[ArxFusion]

Also please put back the restore faces checkbox and is it possible to revert the extra networks to buttons rather than the tab like interface?

[catboxanon]

Face restoration can be added back in Settings -> User interface -> Options in main UI (search for face_restoration).

The extra networks redesign is intentional and there is no way to revert back.

[Kadah]

I greatly prefer the tabbed extra networks. I had the initial PR merged in to my local fork since it was started.

Additional to change in location, its a lot more responsive now, no more load loading times on first open.

What I did not care for was the new much larger thumbnails, but most of UI bugs associated with setting them to be smaller are fixed. The text is still too large and moderately long filenames will fill the entire thumbnail element.

[CNMagicturtle]

TypeError: Expected state_dict to be dict-like, got <class 'NoneType'>.
version: v1.6.0-RC-4-ge004384e  •  python: 3.10.11  •  torch: 2.0.1+cu118  •  xformers: N/A  •  gradio: 3.41.0  •  checkpoint: e6bb9ea85b

[peterpernhofer]

Switched to release_canditate. Will work with it the next few days and let you know, when I notice bugs, etc...

I have a request please:
Is it possible to extend the API functions txt2img and img2img: Please make it possible to pass the whole generation-strings (which are stored in png-chunks/jpeg-exif) in the json prompt parameter. The API should parse this parameters exactly as the Web-UI button "Read generation parameters from prompt" does.
Or is there another possibility to parse those generation-params and fill out a txt2img/img2img payload ?

Thanks in advance :-)

[AUTOMATIC1111]

The way things currently are, it's only possible to read generation parameters into UI in an organized way, which makes it unsuitable for the API. I want to add support for this, but it's not going to be in 1.6.0.

[peterpernhofer]

Thank you anyway. If it works out, that would be very nice :-)

[ArxFusion]

The interrupt button during the generation phase is slightly cropped to the right when using the mobile interface.

[vKireina]

Getting this error when trying to start up the webui, I have 16gb Vram on a Radeon 6900 XT

[silveroxides]

The changes to styles in main tab makes no sense. Why remove/move the button that saves current prompt as style(I can't find it anywhere)? Why put Apply Style to Prompt button in a separate popup box(This only creates an additional action in order to apply styles to prompt and has no beneficial function when compared to previous layout? Looks like wasted space too. Where is the reasoning behind it?

[catboxanon]

See the discussion starting from here: #3062 (comment)

The reasoning is it's now a separate component in the code, so it makes it more difficult to abstract the buttons to live elsewhere in the UI. I want to fix this at some point as well but it's a bit low priority for me currently. PRs are welcome.

[silveroxides]

So presumably the drop down will be removed as well? I noticed it did not load the correct styles until I had first switched to the Style Editor tab and back.
I understand the reasoning from a coding perspective.
But from end user / accessibility perspective it is nothing but a backwards leap. Especially considering the style editor has 'new row' at bottom, only support deletion through keyboard input, Same with the copy/paste and move functions.
Not accessibility or mobile friendly at all. Better implementation would be at right side of row when selected arrows for up/down as well as a trashcan for deletion of entire row appears(confirmation prompt on trashcan interaction), on cell selection three buttons as well but within the selected cell, Copy, Paste, Delete/clear.

[skittlz444]

I'm pretty sure mobile view is not an officially supported view and only works because contributors try and make things responsive. Also for applying styles to prompt, you don't need to send the style text to the prompt box for it to work, you only need to select it from the drop down and they apply.

I think the change is a great one since before there was no good way to view and modify styles at all, only blind overwrite, or append to prompt, or go into the csv file. You can always make changes yourself to add buttons and submit a PR, but these guys are super busy.

[bklstheman]

This change doesnt make any sense. I used styles to help add to the prompt then make some adjustments from there. This new flow is really confusing. I'm going to look to revert away from 1.6 because of this.

[FGhrawi]

Is there a way to make the refiner work before hiresfix ?
Also, any chance we can also set a seperate CFG in the hiresfix section now ?

[dhwz]

Is there a way to make the refiner work before hiresfix ?

Would that make any sense? All details added by the refiner would get lost when applying hires fix afterwards.

[ds1straightup]

I can't figure out how to try out this RC version via colab. Please advise!

[skittlz444]

Can you provide more details as to the output in Colab? I haven't tested it yet, but looking at the versions that were bumped, there aren't many things that should be breaking the Colab build. Also important to note that Colab isn't officially supported by the repo, and is mostly maintained by other repos that specialise in running things via Jupiter notebooks.

My first question would be, did you get version 1.5.2 running in Colab successfully and what were your upgrade steps?

[ds1straightup]

It upgrades to the latest master automatically I’m currently on 1.5.2. I’ve tried changing parts of the automatic1111 install code from master to dev to get it to install from the dev branch but everytime I load it I still have the master release instead. If it helps I’m using thelastben’s Colab notebook for fast diffusion

[Autantpourmoi]

hello , i'm not very good with terminal on macosx , but is it possible to direct this RC installation to the directory as the main install specially for models, embeddings, lora, vae to avoid to clone everything on the HD. Thanks in advance

[skittlz444]

Wait, is the line in your webui-user.sh export COMMANDLINE_ARGS="--ckpt-dir "/Users/lucvaligny/stable-diffusion-webui/models/Stable-diffusion" or export COMMANDLINE_ARGS="--ckpt-dir /Users/lucvaligny/stable-diffusion-webui/models/Stable-diffusion"?

[elistys]

"--ckpt-dir /Volumes/[your Volumes name]/Users/lucvaligny/stable-diffusion-webui/models/Stable-diffusion"

Replace [your Volumes name] with your Mac's volume name.

[Autantpourmoi]

ok thanks guys , will try that

[Autantpourmoi]

skittlz444 the line is COMMANDLINE_ARGS="--ckpt-dir /Users/lucvaligny/stable-diffusion-webui/models/Stable-diffusion""--embeddings-dir /Users/lucvaligny/stable-diffusion-webui/embeddings""--vae-dir /Users/lucvaligny/stable-diffusion-webui/models/VAE""--lora-dir /Users/lucvaligny/stable-diffusion-webui/models/Lora"

[skittlz444]

Ah, don't put quotes between command line arguments, only at the very start and end. Separate them all with spaces instead

[Renaldas111]

Anybody noticed, inpainting models generate extremely dark images? Are you able to generate normal models with inpainting models? v2.1 inpainting, looks, work normally, but v1.5 based give me almost black images.

[AUTOMATIC1111]

I was able to reproduce this and a fix is pushed. 1.6.0 txt2img with inpainting models now produces the same pics as 1.5.2.

[RayHell]

This issue need more visibility.
#12810
If not addressed, 1.6.0 refiner will be broken on release.

[AUTOMATIC1111]

Latest commits makes it possible to choose how to combine refiner with hires fix: either run it on first pass, or on second, or on both -- Hires fix: which pass to enable refiner for setting, on Stable Diffusion page.

[userhomer]

can request an option to turn off controlnet in hires fix pass

[catboxanon]

@userhomer That's unrelated to this repo, but more notably it's been a feature for over a month. Mikubill/sd-webui-controlnet#1832

[suonnon]

Every extensions for CivitAi are not fully compatible with this latest update :( Some loras for SDXL did not appears in Loras tab when those extensions are on :(

[dhwz]

Check the extra networks settings, you may have hidden Lora which are in an "unknown" state, you can set the correct SD version for each Lora in the metadata editor.

[suonnon]

@dhwz No, all of them are in public and visible folder, full permission.
After disable some extensions for CivitAi (for example: CivitAi Helper, CivitAi Shortcut, etc), my SDXL Loras appears again.

[Adzido]

When is the anticipated release date for this version?

[AUTOMATIC1111]

Likely tomorrow. I want to wait at least a week to give time to extension devs to adapt to changes.

[pachencostoner]

Hello to all the developers who made this update possible, thank you!
But I would love to know what happened to Deforum and Infinity zoom?? I haven't seen it anymore since E updated my SD to v1.6 Can someone please help me, I tried to install it but can't. Aid?

[skittlz444]

Most likely those extensions will need an update, deforum at least I know is currently not working with 1.6.0 and the developer of the extension will need to update it, please request it in their repo.

[AlanRexPrime]

Hello. What i have to change in settings to get back option of sending images from txt2img to img2img,extras and vice versa, etc. ? Or is that option gone ?

[skittlz444]

Those options are still there! Just as icons without text, hover over the icons to see what each do :)

[AlanRexPrime]

Yoooo, thanks bro. I was hovering over them, but stopped at third one thinking this ain't it 😂

[f-rank]

Hello to all the developers who made this update possible, thank you! But I would love to know what happened to Deforum and Infinity zoom?? I haven't seen it anymore since E updated my SD to v1.6 Can someone please help me, I tried to install it but can't. Aid?

Deforum team fixed it.

[belenkoksas]

I can see a warning. Is there any timeline for the controller on A1111? 😊

I saw somewhere that the fix will be available on September 5, 2023.
maybe i missed some workaround to run contrnet on SDXL...:((

[sam43b]

thank you for this update , please add save & apply style to the front , plz plz plz

[satangel2222]

may i know why i change Default multiplier for extra networks or Card text scale (1 = original size) also donest work after i click apply setting, i just want to make the lora preview thumnail size smaller. could tell me how to do it in 1.60

[hendrix0420]

This is my setting and the preview cards get smaller. You can set the height and width in pixels.