[Bug]:"Hires Value Prompt" is not correctly recognized and replaced by "Prompt S/R"
Is there an existing issue for this?
- [X] I have searched the existing issues and checked the recent builds/commits
What happened?
I configure the parameters in "ui-config" so that I can use them quickly, such as "/Prompt/visible" and "Prompt S/R".
The default configuration will generate an XYZ grid, but the part in "Hires Value Prompt" is not correctly recognized and replaced by "Prompt S/R"
Steps to reproduce the problem
- set txt2img/Prompt/value and xyz_grid.py/txt2img/X values/value
- use hires.fix
- generate
What should have happened?
"Hires Value Prompt" should not be loaded in "/Prompt/visible" in "ui-config.json" or "Prompt S/R" correctly recognizes and replaces
Commit where the problem happens
1.3.0
What Python version are you running on ?
Python 3.10.x
What platforms do you use to access the UI ?
Windows
What device are you running WebUI on?
Nvidia GPUs (RTX 20 above)
What browsers do you use to access the UI ?
Microsoft Edge
Command Line Arguments
--no-half-vae --api --xformers --opt-channelslast --port 7860 --listen --enable-insecure-extension-access --no-gradio-queue --update-check
List of extensions
a1111-sd-webui-locon a1111-sd-webui-tagcomplete multidiffusion-upscaler-for-automatic1111 sd-dynamic-prompts sd-webui-model-converter sd-webui-supermerger sd-webui-tagger stable-diffusion-webui-dataset-tag-editor stable-diffusion-webui-images-browser stable-diffusion-webui-model-toolkit
Console logs
venv "I:\stable-diffusion-webui\venv\Scripts\Python.exe"
Python 3.10.11 (tags/v3.10.11:7d4cc5a, Apr 5 2023, 00:38:17) [MSC v.1929 64 bit (AMD64)]
Version: v1.3.0
Commit hash: 20ae71faa8ef035c31aa3a410b707d792c8203a3
Installing requirements
Installing sd-dynamic-prompts requirements.txt
You are up to date with the most recent release.
Launching Web UI with arguments: --no-half-vae --api --xformers --opt-channelslast --port 7860 --listen --enable-insecure-extension-access --no-gradio-queue --update-check
Additional Network extension not installed, Only hijack built-in lora
LoCon Extension hijack built-in lora successfully
Loading weights [e81f6bbde2] from I:\stable-diffusion-webui\models\Stable-diffusion\TangorMix_fix1.safetensors
Creating model from config: I:\stable-diffusion-webui\configs\v1-inference.yaml
LatentDiffusion: Running in eps-prediction mode
DiffusionWrapper has 859.52 M params.
Loading VAE weights specified in settings: I:\stable-diffusion-webui\models\VAE\nai_vae.safetensors
Applying optimization: xformers... done.
Textual inversion embeddings loaded(0):
Model loaded in 4.1s (load weights from disk: 0.5s, create model: 0.3s, apply weights to model: 1.1s, apply channels_last: 0.4s, apply half(): 0.5s, load VAE: 0.1s, move model to device: 0.4s, load textual inversion embeddings: 0.7s).
Running on local URL: http://0.0.0.0:7860
To create a public link, set `share=True` in `launch()`.
Startup time: 10.1s (import torch: 1.3s, import gradio: 0.9s, import ldm: 0.3s, other imports: 0.7s, load scripts: 1.5s, create ui: 1.0s, gradio launch: 4.2s).
X/Y/Z plot will create 3 images on 1 3x1 grid. (Total steps to process: 120)
I:\stable-diffusion-webui\models\hypernetworks\noyutest5.pt
Loading hypernetwork noyutest5t/s]
[1.0, 2.0, 2.0, 1.0]
Activation function is swish
Weight initialization is Normal
Layer norm is set to False
Dropout usage is set to True
Activate last layer is set to False
Dropout structure is set to [0.0, 0.3, 0.3, 0.0]
No saved optimizer exists in checkpoint
100%|██████████████████████████████████████████████████████████████████████████████████| 24/24 [00:04<00:00, 5.41it/s]
100%|██████████████████████████████████████████████████████████████████████████████████| 16/16 [00:04<00:00, 3.29it/s]
[Tiled VAE]: the input size is tiny and unnecessary to tile. | 40/120 [00:10<00:24, 3.32it/s]
I:\stable-diffusion-webui\models\hypernetworks\nibiirotest11.pt
Loading hypernetwork nibiirotest11
[1.0, 2.0, 2.0, 1.0]
Activation function is swish
Weight initialization is Normal
Layer norm is set to False
Dropout usage is set to True
Activate last layer is set to False
Dropout structure is set to [0.0, 0.3, 0.3, 0.0]
No saved optimizer exists in checkpoint
100%|██████████████████████████████████████████████████████████████████████████████████| 24/24 [00:02<00:00, 10.40it/s]
I:\stable-diffusion-webui\models\hypernetworks\noyutest5.pt | 64/120 [00:14<00:05, 10.89it/s]
Loading hypernetwork noyutest5
[1.0, 2.0, 2.0, 1.0]
Activation function is swish
Weight initialization is Normal
Layer norm is set to False
Dropout usage is set to True
Activate last layer is set to False
Dropout structure is set to [0.0, 0.3, 0.3, 0.0]
No saved optimizer exists in checkpoint
100%|██████████████████████████████████████████████████████████████████████████████████| 16/16 [00:04<00:00, 3.31it/s]
[Tiled VAE]: the input size is tiny and unnecessary to tile.████▎ | 80/120 [00:19<00:12, 3.32it/s]
100%|██████████████████████████████████████████████████████████████████████████████████| 24/24 [00:01<00:00, 14.28it/s]
I:\stable-diffusion-webui\models\hypernetworks\noyutest5.pt████████████████▉ | 103/120 [00:22<00:01, 13.26it/s]
Loading hypernetwork noyutest5
[1.0, 2.0, 2.0, 1.0]
Activation function is swish
Weight initialization is Normal
Layer norm is set to False
Dropout usage is set to True
Activate last layer is set to False
Dropout structure is set to [0.0, 0.3, 0.3, 0.0]
No saved optimizer exists in checkpoint
100%|██████████████████████████████████████████████████████████████████████████████████| 16/16 [00:04<00:00, 3.32it/s]
[Tiled VAE]: the input size is tiny and unnecessary to tile.█████████████████████████| 120/120 [00:27<00:00, 3.34it/s]
Total progress: 100%|████████████████████████████████████████████████████████████████| 120/120 [00:30<00:00, 4.00it/s]
Total progress: 100%|████████████████████████████████████████████████████████████████| 120/120 [00:30<00:00, 3.34it/s]
Additional information
No response
apparently there is currently a lack of support for Hires Prompt in xyz
adding support is easy
but how to add support elegantly
@w-e-w I was about to open a new issue regarding Hi-Res fix, but the issue has much to do with what is reported here... ...except, X/Y/Z just seems to be another aspect of the bug.
Like dorioku I have preconfigured prompt for txt-2-img.
With the default A1111 settings (Settings > Hi-Res Sampler // Hi-Res prompt = both UNCHECKED), what happens is that the Hi-Res prompt (user cannot see this) just defaults to whatever is predefined for the txt2img prompt in config file.
I was going nuts trying to figure out why my images were changing dramatically during HiRes since updating to new SD version. I only figured it out once reviewing my image metadata, looking at those hidden settings, and figuring this out.
Let me know if I should open a new Bug for this. I feel like this bug report here only describes another side effect of the more significant bug.
我正要打开一个关于高分辨率修复的新问题,但这个问题与这里报告的内容有很大关系...... 。除了,X / Y/ Z似乎只是错误的另一个方面。
像dorioku一样,我已经预配置了txt-2-img的提示。
使用默认的 A1111 设置(高分辨率采样器>设置 // 高分辨率提示符 = 两个未选中),发生的情况是高分辨率提示符(用户看不到此提示符)仅默认为配置文件中为 txt2img 提示符预定义的任何内容。
我快疯了,试图弄清楚为什么自从更新到新的SD版本以来,我的图像在HiRes期间发生了巨大变化。我只是在查看我的图像元数据、查看那些隐藏的设置并弄清楚这一点后才弄清楚。
让我知道我是否应该为此打开一个新的错误。我觉得这里的这个错误报告只描述了更重要的错误的另一个副作用。
是的 我之前也在寻找为什么导致这种情况 直到我在设置中发现了 "Hires fix: show hires sampler selection(requires restart) Hires fix: show hires prompt and negative prompt(requires restart)" Their default value is not enabled
Indeed.
Any person who edits user-config.json and adds a prompt (like we both did), will suddenly have bad results from Hi-Res fix.
apparently there is currently a lack of support for Hires Prompt in xyz adding support is easy but how to add support elegantly
didn't explain well in my initial response responce there was a rework on how prompt is handled for hires-fix that effective XYZ S/R
this can be easily fixe but I want to fix it in a more user-friendly way while adding less clutter while providing more control to the user
Frankly I don't know which commit resolved this but according to voldy it's been fixed, and for a while now. https://github.com/AUTOMATIC1111/stable-diffusion-webui/issues/10892#issuecomment-1571466191
Closing.