haoming02 / sd-webui-vectorscope-cc

An Extension for Automatic1111 Webui that performs Offset Noise* natively

Profile

• 👋 Hi, I’m @Haoming02
• 💼 Professional Unity Developer
• 🤖 AI Art Enthusiast
• 📫 Contact: [email protected]

Socials

• 🎥 YouTube
• 🧰 CivitAI
• 🖼️ Pixiv
• ☕ Ko-fi
  • Fund my GPU~

Accumulated Stars ✨

sd-webui-vectorscope-cc's Issues

working with different samplers

I see that the extensions works differently with different samplers.
Is this the expected behavior?

A 30 year old tall smiling [Leona Lewis | Laura Vandervoort | Katheryn Winnick] Negative prompt: bad or low quality, low resolution Steps: 24, Sampler: Euler, CFG scale: 4, Seed: 2838915928, Size: 512x768, Model hash: c35822169d, Model: YACAM-SR-33614, RNG: CPU, Vec CC Enabled: True, Vec CC Alt: False, Vec CC Brightness: -5, Vec CC Contrast: 5, Vec CC Saturation: 1, Vec CC R: 0, Vec CC G: 0, Vec CC B: 0, Vec CC Noise: Straight Abs., Vec CC Proc HrF: False, Vec CC Proc Ade: False, Vec CC Scaling: Flat, Vec CC Version: v2.0.1, MuDDetailer use prompt edit: False, MuDDetailer use prompt edit b: False, MuDDetailer model a: yolo/face_yolov8s.pt [2e157766], MuDDetailer conf a: 30, MuDDetailer max detection a: 1, MuDDetailer dilation a: 4, MuDDetailer offset x a: 0, MuDDetailer offset y a: 0, MuDDetailer mask blur: 4, MuDDetailer denoising: 0.4, MuDDetailer inpaint full: True, MuDDetailer inpaint padding: 32, MuDDetailer inpaint width: 0, MuDDetailer inpaint height: 0, MuDDetailer CFG scale: 0, MuDDetailer steps: 0, MuDDetailer noise multiplier: 0, MuDDetailer detection a: "A-face 0.8911,141,104,326,360", Script: X/Y/Z plot, X Type: Sampler, X Values: "Euler,DPM++ 2M SDE Karras,DPM++ 3M SDE Karras,Restart,UniPC", Version: v1.7.0

Adetailer

Is there a way to completely disable it for adetailer? I noticed it was making images very dark, particularly in the 2nd adetailer model.

[Feature Request]Add a color wheel/color ring/picker or something like that.

Is it possible to add a color wheel/ring/picker or something like that to replace the original RGB sliders? This would allow visualization of the expected offset hue, rather than an abstract RGB slider.

[Feature Request] Randomize all

This extension gives vast generation diversity. Would be cool to randomize contrast, brightness, etc, so that I don't have to tweak them myself between generations. Would be even better if it works for batches, similar to dynamic prompts. I think the most obvious choice for randomization is sampling a normal distribution for each parameter.

Doesn't Work with Selected Few LoRA

When trying to use with a LoRA model, the effect doesn't feel apparent...
Can't really generate a dark image with a specific character...

AttributeError: hr_pass / line 192, in postprocess_image / del p.hr_pass

I get error messages when generating images since updating to the latest version of this extension ([f3ffc5f]) under the following conditions:

Vectorscope CC extension OFF = error message below
Vectorscope CC extension ON + Hires Fix OFF = error message below
Vectorscope CC extension ON + Hires Fix ON + Vectorscope CC Process Hires. fix ON = error message below
Vectorscope CC extension ON + Hires Fix ON + Vectorscope CC Process Hires. fix OFF = OK, no error

And here is the error message logged on the WebUI Command Prompt window

*** Error running postprocess_image: C:\stable-diffusion-webui\extensions\sd-webui-vectorscope-cc\scripts\cc.py.94it/s]
    Traceback (most recent call last):
      File "C:\stable-diffusion-webui\modules\scripts.py", line 514, in postprocess_image
        script.postprocess_image(p, pp, *script_args)
      File "C:\stable-diffusion-webui\extensions\sd-webui-vectorscope-cc\scripts\cc.py", line 192, in postprocess_image
        del p.hr_pass
    AttributeError: hr_pass

System info:
version: v1.3.2-241-g59419bd6  •  python: 3.10.6  •  torch: 2.0.0+cu118  •  xformers: N/A  •  gradio: 3.32.0  •  checkpoint: f04fdb4a9e

Error installing it in A1111 at Google Colab

Hi,

I am using A1111 using the last Ben´s colab, through an Google Colab paid account and I am unable to install this extension.
It doesn´t matter if I install it from URL, or from the list of available extensions. In both cases I get this message on the A1111 GUI´s :

GitCommandError: Cmd('git') failed due to: exit code(255) cmdline: git clone -v --filter=blob:none https://github.com/Haoming02/sd-webui-vectorscope-cc.git /content/gdrive/MyDrive/sd/stable-diffusion-webui/tmp/sd-webui-vectorscope-cc stderr: 'Cloning into '/content/gdrive/MyDrive/sd/stable-diffusion-webui/tmp/sd-webui-vectorscope-cc'... POST git-upload-pack (175 bytes) POST git-upload-pack (287 bytes) Filtering content: 71% (10/14) Filtering content: 78% (11/14) Filtering content: 85% (12/14) Filtering content: 92% (13/14) Filtering content: 100% (14/14) Filtering content: 100% (14/14), 16.74 MiB | 14.46 MiB/s, done. fatal: cannot exec '/content/gdrive/MyDrive/sd/stable-diffusion-webui/tmp/sd-webui-vectorscope-cc/.git/hooks/post-checkout': Permission denied '

I tried to install other extensions, and I didn´t have any problem.

Batch support

Really good extension, works perfectly, only issue currently is it only applies to the first image in a batch, though I'm not sure if it's only my side.

First image with blue and saturation cranked up as an example.

Request: Save advanced settings when you save a style

Currently options made in advanced settings aren't saved when you save a style.

Auto ticked when started with A1111's WebUI 1.8

It now gets automatically ticked with the starting of the WebUI, and the 'Brightness' will be set to '-1' instead of '0'. Is that a bug?

presets already downloaded in fold model/lut,but cannot load

refresh style, it doesn't work, presets is empty nothing to choose,,,,while all the cubes have already been downloaded just cannot show

Wrong coworking with MuDDetailer

Hi! Could you please add support for MuDDetailer as it's already done for Adetailer?
https://github.com/wkpark/uddetailer

A 30 year old tall smiling [Leona Lewis | Laura Vandervoort | Katheryn Winnick] Negative prompt: bad or low quality, low resolution Steps: 24, Sampler: DPM++ 3M SDE Karras, CFG scale: 4, Seed: 2838915928, Size: 512x512, Model hash: c35822169d, Model: YACAM-SR-33614, RNG: CPU, Vec CC Enabled: True, Vec CC Alt: False, Vec CC Brightness: -5, Vec CC Contrast: 5, Vec CC Saturation: 1, Vec CC R: 0, Vec CC G: 0, Vec CC B: 0, Vec CC Noise: Multi-Res, Vec CC Proc HrF: False, Vec CC Proc Ade: False, Vec CC Scaling: Cos, Vec CC Version: v2.0.1, Eta: 0.5, MuDDetailer use prompt edit: False, MuDDetailer use prompt edit b: False, MuDDetailer model a: yolo/face_yolov8s.pt [2e157766], MuDDetailer conf a: 30, MuDDetailer max detection a: 1, MuDDetailer dilation a: 4, MuDDetailer offset x a: 0, MuDDetailer offset y a: 0, MuDDetailer mask blur: 4, MuDDetailer denoising: 0.4, MuDDetailer inpaint full: True, MuDDetailer inpaint padding: 32, MuDDetailer inpaint width: 0, MuDDetailer inpaint height: 0, MuDDetailer CFG scale: 0, MuDDetailer steps: 0, MuDDetailer noise multiplier: 0, MuDDetailer detection a: "A-face 0.895,215,89,352,279", Version: v1.7.0

[Feature Request] Save to preset

The new randomize feature is pretty nice and can give some very interesting results, it would be cool if there was an option to save those values to presets to preserve them for future use.

Notification: scheduled to be added to webui extension index if no objection is raised

Hi @Haoming02 I'm one of the maintenance of webui extension Index
we'd like to add your extension https://github.com/Haoming02/sd-webui-vectorscope-cc to the extension index
we would like to confirm with you before adding it
see AUTOMATIC1111/stable-diffusion-webui-extensions#169

scheduled to be added after a week if we did not receive any negative response from you

[Feature Request] 32-bit HDR Output Workflow

I managed to use this extension to create a passable approximation of a high-dynamic range image in 32 bit per channel. I must use photoshop to do some of the steps, but most of the hard work is done with this extension.
I would like to describe my workflow and have your input but I am not sure if this "issue" tab is the right place to do it.
There are also a few feature requests I have in mind to help me with this HDR generation process, but let's describe my current workflow prototype first and see what you think.

I'm also working on panoramic HDRI environment maps, which are quite useful for real-time 3d engines. More on that soon !

Finally, THANK YOU SO MUCH for making this extension. I was already a fan, and I had already tried to do what I just did today with an earlier version of your extension, but it looks like the latest version has the features that were missing to make this whole HDR image generation work. You made this possible by creating and sharing this extension, and I can't thank you enough for that.

[Feature Request]Multi-Resolution-Noise

https://wandb.ai/johnowhitaker/multires_noise/reports/Multi-Resolution-Noise-for-Diffusion-Model-Training--VmlldzozNjYyOTU2

Standard Implementation
kohya-ss/sd-scripts#489

[QUESTION] Maybe an issue ? Contrast slider

Hello, first off, I freakin love this extension! Lots of potential.

Second, I think the contrast slider has some issues because I've tried with all the settings with XYZ grids, and the contrast doesn't seem to have any major impact. Which is a shame because I would love to choose how much contrast the output images have by either lowering it to make the image more flat, or increasing it for more depth.

How come ?

how to disable the method in Noise Settings?

like this:

`Euler` sampler only works with `Alt.` method

The default behavior (not enabling Alt. (Stronger Effects)) doesn't seem to have any affect for me currently.

Below are attempts to reproduce something like that shown in the sample images. Images have infotext included. I don't have the SDXL lightning LoRA in use but I assume that shouldn't make a difference.

Disabled

Enabled

Enabled (with "Alt." method)

Tested on Forge and upstream A1111 dev branch.

lllyasviel/stable-diffusion-webui-forge@b9705c5
AUTOMATIC1111/stable-diffusion-webui@7d1368c

Infotext paste support

Title. Currently this isn't possible because there's a couple problems that need to be addressed:

1. Because the generation params are currently dumped into a single string, a parser would need to be written using the on_infotext_pasted callback to retrieve the values. These are intended to be stored under individual keys.
2. Furthermore, because the generation params are added under a single key with a version number that is variable (Vec. CC [{VERSION}]), it would prevent the ability to read params from infotext from a different version of the extension. Ideally this would just be stored under a separate generation params key.