AUTOMATIC1111
[Bug]: Stable Diffusion stuck on waiting when clicking generate on tab img2img
Is there an existing issue for this?
- [X] I have searched the existing issues and checked the recent builds/commits
What happened?
I have a problem when my sd updates to the latest version. I notice the change arrangement of the "Sampling method" when the "dpm++ 2M karras, dpm++ SDE karras, dpm++ 2M SDE Exponential, dpm++ 2M SDE Karras" lines are moved above the "euler a" line. In tab txt2img, the sampling methods work normal, but tab img2img I get stuck on waiting when use certain specific methods, terminal not report error. I remember the previous version I could use these methods "dpm++ 2M karras, dpm++ SDE karras, dpm++ 2M SDE Karras". plus that the SD will stuck on waiting when generate the second time. I also tried reinstalling the SD, tried a few ways that I found out, but it still didn't fix. I haven't tried all the sampling methods yet, see what is bug and what is not.
Steps to reproduce the problem
- Go to .... img2img
- Drop an image here
- Choose a sampling method (bug in DPM++2M Karras, DPM++SDE Karras... but euler a, euler, LMS... still work normal)
- Press Generate
- Waiting....
What should have happened?
Images should appear in the display box
Sysinfo
What browsers do you use to access the UI ?
Microsoft Edge
Console logs
E:\Stable Diffusion\webui-automatic1111\stable-diffusion-webui-directml>git pull
Already up to date.
venv "E:\Stable Diffusion\webui-automatic1111\stable-diffusion-webui-directml\venv\Scripts\Python.exe"
fatal: No names found, cannot describe anything.
Python 3.10.6 | packaged by conda-forge | (main, Oct 24 2022, 16:02:16) [MSC v.1916 64 bit (AMD64)]
Version: 1.6.0
Commit hash: 39e6d26f3b66805677cc3b5c2bc3bd624c347608
Launching Web UI with arguments: --backend directml --no-half --precision full --opt-sub-quad-attention --opt-split-attention-v1 --disable-nan-check --medvram
no module 'xformers'. Processing without...
no module 'xformers'. Processing without...
No module 'xformers'. Proceeding without it.
Loading weights [6ce0161689] from E:\Stable Diffusion\webui-automatic1111\stable-diffusion-webui-directml\models\Stable-diffusion\v1-5-pruned-emaonly.safetensors
E:\Stable Diffusion\webui-automatic1111\stable-diffusion-webui-directml\extensions\sd-extension-system-info\scripts\system-info.py:451: GradioDeprecationWarning: The `style` method is deprecated. Please set these arguments in the constructor instead.
refresh_quick_btn = gr.Button('Refresh state', elem_id = 'system_info_tab_refresh_btn', visible = False).style() # quick refresh is used from js interval
E:\Stable Diffusion\webui-automatic1111\stable-diffusion-webui-directml\extensions\sd-extension-system-info\scripts\system-info.py:452: GradioDeprecationWarning: The `style` method is deprecated. Please set these arguments in the constructor instead.
refresh_full_btn = gr.Button('Refresh data', elem_id = 'system_info_tab_refresh_full_btn', variant='primary').style()
E:\Stable Diffusion\webui-automatic1111\stable-diffusion-webui-directml\extensions\sd-extension-system-info\scripts\system-info.py:504: GradioDeprecationWarning: The `style` method is deprecated. Please set these arguments in the constructor instead.
bench_run_btn = gr.Button('Run benchmark', elem_id = 'system_info_tab_benchmark_btn', variant='primary').style()
E:\Stable Diffusion\webui-automatic1111\stable-diffusion-webui-directml\extensions\sd-extension-system-info\scripts\system-info.py:506: GradioDeprecationWarning: The `style` method is deprecated. Please set these arguments in the constructor instead.
bench_submit_btn = gr.Button('Submit results', elem_id = 'system_info_tab_submit_btn', variant='primary').style()
E:\Stable Diffusion\webui-automatic1111\stable-diffusion-webui-directml\extensions\sd-extension-system-info\scripts\system-info.py:515: GradioDeprecationWarning: The `style` method is deprecated. Please set these arguments in the constructor instead.
refresh_bench_btn = gr.Button('Refresh bench', elem_id = 'system_info_tab_refresh_bench_btn', visible = False).style(full_width = False) # quick refresh is used from js interval
E:\Stable Diffusion\webui-automatic1111\stable-diffusion-webui-directml\extensions\sd-extension-system-info\scripts\system-info.py:515: GradioDeprecationWarning: Use `scale` in place of full_width in the constructor. scale=1 will make the button expand, whereas 0 will not.
refresh_bench_btn = gr.Button('Refresh bench', elem_id = 'system_info_tab_refresh_bench_btn', visible = False).style(full_width = False) # quick refresh is used from js interval
Running on local URL: http://127.0.0.1:7860
To create a public link, set `share=True` in `launch()`.
Creating model from config: E:\Stable Diffusion\webui-automatic1111\stable-diffusion-webui-directml\configs\v1-inference.yaml
Startup time: 9.0s (prepare environment: 0.3s, import torch: 3.0s, import gradio: 0.8s, setup paths: 0.7s, initialize shared: 1.1s, other imports: 0.4s, load scripts: 1.1s, create ui: 0.5s, gradio launch: 1.1s).
Applying attention optimization: sub-quadratic... done.
Model loaded in 2.3s (load weights from disk: 0.8s, create model: 0.3s, apply weights to model: 0.9s, calculate empty prompt: 0.2s).
6%|████ | 1/16 [00:01<00:22, 1.53s/it]
100%|████████████████████████████████████████████████████████████████| 16/16 [00:11<00:00, 1.38it/s]
100%|████████████████████████████████████████████████████████████████| 16/16 [00:10<00:00, 1.52it/s]
Total progress: 100%|████████████████████████████████████████████████| 32/32 [00:23<00:00, 1.35it/s]
Total progress: 100%|████████████████████████████████████████████████| 32/32 [00:23<00:00, 1.60it/s]
Additional information
https://github.com/AUTOMATIC1111/stable-diffusion-webui/assets/144250978/63644921-7c05-4fab-a75c-006567e87c96
I had a similar issue. Before updating I could generate img2img at 960 * 1440, and txt2img at 512 * 768 with Hires at 1.87. That was my pixel count limit on a laptop with a 1660Ti. After the update I couldn't work at that size because it just hangs forever. Made a clean install of 1.5.2 and same problem. What worked at the end was making a clean installation of 1.6 and now I can generate again at my hardware limit. I also played around with Nvidia drivers so that could also have had an effect. GL.
