AUTOMATIC1111 / stable-diffusion-webui

Stable Diffusion web UI
GNU Affero General Public License v3.0
135.35k stars 25.85k forks source link

[Bug]: SD generates black images. #13722

Open scorobetele opened 8 months ago

scorobetele commented 8 months ago

Is there an existing issue for this?

What happened?

It generates black images. I have been trying to fix it for hours, but to no avail.

Steps to reproduce the problem

  1. click on "Generate"
  2. wait
  3. cry because it generated a black image despite trying for several hours to fix it

What should have happened?

It should have generated an image, that is not black.

Sysinfo

sysinfo-2023-10-22-09-48.txt

What browsers do you use to access the UI ?

Brave

Console logs

venv "C:\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: v1.6.0
Commit hash: 5ef669de080814067961f28357256e8fe27544f4
Launching Web UI with arguments: --xformers --medvram --autolaunch --no-half-vae --no-half --precision full --lowvram --always-batch-cond-uncond --upcast-sampling --api --listen --disable-nan-check --skip-torch-cuda-test
Loading weights [f2139304e6] from C:\stable-diffusion\webui\models\Stable-diffusion\MedievalArmor_v2.safetensors
Creating model from config: C:\stable-diffusion\webui\configs\v1-inference.yaml
Running on local URL:  http://0.0.0.0:7860
Applying attention optimization: xformers... done.
Model loaded in 4.3s (load weights from disk: 0.8s, create model: 1.0s, calculate empty prompt: 2.4s).

To create a public link, set `share=True` in `launch()`.
Startup time: 15.9s (prepare environment: 0.4s, import torch: 4.6s, import gradio: 1.3s, setup paths: 1.1s, initialize shared: 0.3s, other imports: 0.8s, setup codeformer: 0.2s, load scripts: 1.9s, create ui: 0.7s, gradio launch: 4.6s, add APIs: 0.1s).
100%|██████████████████████████████████████████████████████████████████████████████████| 20/20 [01:00<00:00,  3.01s/it]
Total progress: 100%|██████████████████████████████████████████████████████████████████| 20/20 [00:59<00:00,  2.97s/it]
Total progress: 100%|██████████████████████████████████████████████████████████████████| 20/20 [00:59<00:00,  2.67s/it]

Additional information

I have an Nvidia 1660 Super GPU and I believe this is the source of the problem because it seems that 16xx GPUs have this bug, but unfortunately none of the fixes worked for me.

HG2407 commented 8 months ago

same error here. On using the finetuned safetensors. It either generated most random noise or dark greyish image, even on previously tested community models

byteconcepts commented 7 months ago

Same here

PaulSinghDev commented 1 month ago

Did any of you get this fixed? I'm facing the same issue. No matter what I prompt I just get noise back

PaulSinghDev commented 3 weeks ago

In my case, above, the issue went away when i used --no-half on the M3 mac.