Image generation outputs the picture very slow while watching a youtube video

TheHeimZocker commented 9 months ago

Checklist

[X] The issue exists after disabling all extensions
[X] The issue exists on a clean installation of webui
[ ] The issue is caused by an extension, but I believe it is caused by a bug in the webui
[X] The issue exists in the current version of the webui
[X] The issue has not been reported before recently
[ ] The issue has been reported before but has not been fixed yet

What happened?

As said in the title, the actual output is very slow, the generation of it not. So example: I start the generation, it goes all the way to the 5th step and it just keeps loading for minutes. (while i have a youtube video running on 1080p) I tried lowering the resolution and the problem still occurs. Once i stop the video, it runs completely fine and outputs the picture really fast.

Screenshot_43

Steps to reproduce the problem

Have a youtube video running - start the generation of the picture

What should have happened?

Not slow down the output of the picture

What browsers do you use to access the UI ?

Mozilla Firefox

Sysinfo

sysinfo-2024-01-10-18-08.json

Console logs

venv "G:\stable-diffusion\stable-diffusion-webui\venv\Scripts\Python.exe"
Python 3.10.6 (tags/v3.10.6:9c7b4bd, Aug  1 2022, 21:53:49) [MSC v.1932 64 bit (AMD64)]
Version: v1.7.0
Commit hash: cf2772fab0af5573da775e7437e6acdca424f26e
Launching Web UI with arguments: --api --xformers --api-server-stop
Style database not found: G:\stable-diffusion\stable-diffusion-webui\styles.csv
Loading weights [e686128cf0] from G:\stable-diffusion\stable-diffusion-webui\models\Stable-diffusion\pixelwaveturboExcellent_02.safetensors
Running on local URL:  http://127.0.0.1:7860

To create a public link, set `share=True` in `launch()`.
Creating model from config: G:\stable-diffusion\stable-diffusion-webui\repositories\generative-models\configs\inference\sd_xl_base.yaml
Startup time: 19.6s (prepare environment: 5.0s, import torch: 4.8s, import gradio: 2.0s, setup paths: 2.4s, initialize shared: 0.2s, other imports: 1.5s, setup codeformer: 0.2s, list SD models: 0.2s, load scripts: 1.6s, create ui: 0.6s, gradio launch: 1.2s).
Applying attention optimization: xformers... done.
Model loaded in 10.4s (load weights from disk: 1.4s, create model: 0.7s, apply weights to model: 7.6s, move model to device: 0.1s, load textual inversion embeddings: 0.1s, calculate empty prompt: 0.4s).
100%|████████████████████████████████████████████████████████████████████████████████████| 5/5 [00:07<00:00,  1.52s/it]
========================================================================================== 5/5 [00:06<00:00,  1.38s/it]
A tensor with all NaNs was produced in VAE.
Web UI will now convert VAE into 32-bit float and retry.
To disable this behavior, disable the 'Automatically revert VAE to 32-bit floats' setting.
To always start with 32-bit VAE, use --no-half-vae commandline flag.
==========================================================================================
Total progress: 100%|████████████████████████████████████████████████████████████████████| 5/5 [00:14<00:00,  2.99s/it]
100%|████████████████████████████████████████████████████████████████████████████████████| 5/5 [00:07<00:00,  1.41s/it]
Total progress: 100%|████████████████████████████████████████████████████████████████████| 5/5 [00:11<00:00,  2.29s/it]
100%|████████████████████████████████████████████████████████████████████████████████████| 5/5 [00:07<00:00,  1.48s/it]
Total progress: 100%|████████████████████████████████████████████████████████████████████| 5/5 [00:12<00:00,  2.41s/it]
100%|████████████████████████████████████████████████████████████████████████████████████| 5/5 [00:07<00:00,  1.46s/it]
Total progress: 100%|████████████████████████████████████████████████████████████████████| 5/5 [00:16<00:00,  3.30s/it]
100%|████████████████████████████████████████████████████████████████████████████████████| 5/5 [00:07<00:00,  1.48s/it]
Total progress: 100%|████████████████████████████████████████████████████████████████████| 5/5 [00:11<00:00,  2.39s/it]
100%|████████████████████████████████████████████████████████████████████████████████████| 5/5 [00:07<00:00,  1.48s/it]
Total progress: 100%|████████████████████████████████████████████████████████████████████| 5/5 [00:22<00:00,  4.50s/it]
100%|████████████████████████████████████████████████████████████████████████████████████| 5/5 [00:07<00:00,  1.48s/it]
Total progress: 100%|████████████████████████████████████████████████████████████████████| 5/5 [00:12<00:00,  2.40s/it]
100%|████████████████████████████████████████████████████████████████████████████████████| 5/5 [00:07<00:00,  1.47s/it]
Total progress: 100%|████████████████████████████████████████████████████████████████████| 5/5 [00:14<00:00,  2.83s/it]
100%|████████████████████████████████████████████████████████████████████████████████████| 5/5 [00:07<00:00,  1.47s/it]
Total progress: 100%|████████████████████████████████████████████████████████████████████| 5/5 [00:11<00:00,  2.39s/it]
100%|████████████████████████████████████████████████████████████████████████████████████| 5/5 [00:07<00:00,  1.53s/it]
Total progress: 100%|████████████████████████████████████████████████████████████████████| 5/5 [04:04<00:00, 48.98s/it]
Total progress: 100%|████████████████████████████████████████████████████████████████████| 5/5 [04:04<00:00,  1.49s/it]

Additional information

No response

EclipsedStar0 commented 9 months ago

I (think) this might have to do with the fact that to watch a YouTube video, you're diverting resources from your GPU to watching the video instead of generating the image.

Woisek commented 9 months ago

@TheHeimZocker Guess what is used when watching a video ,,, ? Or playing a game ... ? Or any video output related usage ... ?

TheHeimZocker commented 9 months ago

@EclipsedStar0 @Woisek Once again, the generating of the image itself is going fast. It's the output that takes long, like, the last push if i where to call it. Also, @Woisek thanks for the input, but a video is using video decoding mostly in my experience.

BrickMissle commented 9 months ago

watch youtube on your phone while generating

ipinz commented 9 months ago

Try to disable hardware acceleration in setting Screenshot 2024-01-11 072857

Woisek commented 9 months ago

@EclipsedStar0 @Woisek Once again, the generating of the image itself is going fast. It's the output that takes long, like, the last push if i where to call it.

The last thing is applying the VAE ... the heaviest of this process. What do you expect?

Also, @Woisek thanks for the input, but a video is using video decoding mostly in my experience.

And where do you think does this happen, hm? Hint: It starts with V and ends with RAM ...

AUTOMATIC1111 / stable-diffusion-webui